Studies on Altered Connectivity in the Fragile X Dentate Gyrus
and its Relevance to Autism
Werner Dykstra
Studentnumber: 10591419
Mentors: Ben Schmand (UvA) and Abdeslem El Idrissi (CUNY) No. of words: 6506
Abstract
The current research investigated the functional consequences of altered connectivity in the dentate gyrus (DG) of fmr1-ko mice and its relevance to autism. Five fmr1-ko mice and five control mice were first of all subjected to a Novel Object Recognition (NOR) test, an Acoustic Startle Response (ASR) test and a three chambered social interaction test (TCSIT). After being sacrificed, their brains were prepared for paired-pulse electrophysiological stimulation at the DG perforant path, recordings at the Schaffer collateral and
immunohistochemical analysis with a confocal microscope and a western blot. Frm1-ko mice exhibited impaired NOR, an increased ASR, impaired pre-pulse inhibition, reduced social interest and altered DYRK1A, PSD-95, Gephyrin and GABAA receptor expression in the cerebellum (CB), hippocampus (HC), DG and cortex (CX). Also, increased excitability was observed during electrophysiology. These findings suggest that altered DG connectivity due to absence of FMRP is one of the underlying mechanism of autism and fragile X.
Introduction
Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder characterized by impairments in communication, social interaction and patterns of repetitive or stereotyped behavior (Freitag & Konrad, 2014) not rarely accompanied by intellectual disabilities,
epilepsy and sleep disorders (Yu & Berry-kravis, 2014).The neurobiological mechanisms that underlie ASD are not entirely understood, as they involve multiple, complex genetic
interactions (Willsey & State, 2015) and possibly, environmental conditions (Mandy & Lai, 2016). Fragile X Syndrome (FXS) is a monogenic neurodevelopmental retardation syndrome caused by a trinucleotide repetition (CGG) in the Fragile X Mental Retardation 1 (fmr1) gene on the X-chromosome. This triplet expansion causes transcriptional silencing of this gene, which leads to a loss of or deficiency in the Fragile X Mental Retardation Protein (FMRP) (McDevitt, Gallagher & Reilly, 2015). Behavioral symptoms of FXS include tactile
defensiveness, gaze avoidance, repetitive behaviors, impaired speech, hyper arousal, anxiety and increased susceptibility to seizures. Approximately 25% of males with Fragile X meet the DSM-IV criteria for autism (Harris et al., 2008). The phenotypical parallels between ASD and FXS suggest that the effects of the genetic mutation of FXS converge with the underlying brain mechanisms of ASD. The current research investigated the functional consequences of altered connectivity in the dentate gyrus (DG) of the fragile X mouse brain and its relevance to autism.
Fragile X knock-out (fmr1-KO) mice lack FMRP and its phenotype resembles
elements of FXS, including mildly impaired learning behavior, altered anxiolytical behavior, reduced social interaction, hyperactivity, patterns of repetitive behaviors, increased seizure susceptibility (El Idrissi, Neuwirth & L’Amoreaux, 2010) and spatial learning deficits, which make them a good model for FXS and ASD (Moy & Nadler, 2008). The DG, as part of the hippocampal structures, has been linked to autism for its role in pattern separation of
information (Leutgeb, 2007) and because of its reduced size found in autistic men and boys (Saitoh, Karns & Courchesne, 2001).
Volumetric brain abnormalities are common in both ASD and FXS (Hazlett et al, 2012) and are accompanied by abnormal connectivity (Van der Molen, Stam & Van der Molen, 2014). A common theory of ASD brain connectivity is that of long distance hypo-connectivity and local hypo-connectivity (Courchesne & Pierce, 2005). Cortical hyper-connectivity found within brain regions might be associated with superior acquisition of basic information like sensory perception or even hypersensitivities, the ability to memorize scripts and the focus on details, whereas functional hypo-connectivity between regions such as the association cortex or cortico-limbic pathways might underlie the inability to properly integrate information, understand metaphors and exhibit empathy (El Idrissi, in preparation).
An imbalance of excitatory and inhibitory neural signaling has been suggested to play a crucial role in abnormal brain development seen in autism (Coghlan, 2012) and FXS (El Idrissi et al., 2005) (Van der Molen et al., 2014). Functional and structural MRI studies support the theory of long distance hypo-connectivity in ASD, but not all of them are in line with that of local hyper-connectivity (Vissers, Cohen & Geurts, 2012). These conflicting results may be due to age-related and methodological factors, emphasizing the need for electrophysiology and multi-disciplined studies (Kana, Uddin, Kenet, Chugani & Müller, 2014). Abnormal brain connectivity due to genetically mediated axonal and synaptic abnormalities may be crucial for understanding the neurobiology of ASD (McFadden and Minshew (2013) and FXS (Yang et al., 2016).
Hinton (1991) reported an increase in synaptic spine density and a larger proportion of longer, morphologically immature synaptic spines in people with FXS. Fmr1-KO mice
exhibit elongated dendritic spines and reduced synaptic maturation due to loss of FMRP expression (Pfeiffer & Huber, 2009). The loss of FMRP expression alters the expression of
long-term potentiation (LTP) and long-term depression (LTD). It is understood that LTP is important for the encoding of memory (Cooke, 2006) and LTD seems to play an important role in synapse weakening, suggesting that the collaboration of these mechanisms contributes to learning and memory storage throughout life (Massey & Bashir, 2007). Research on synaptic plasticity in fmr1-KO mice has led to the metabotropic glutamate receptor (mGluR) theory of FXS, based on research on the hippocampal area CA1 (Bear, Huber & Warren, 2004). Bear and colleagues proposed that the loss of FMRP increases signaling via mGluRs, which leads to an increase in LTD. Consequently this is thought to manifest as proportionally less synapse gain and more synapse weakening during brain development resulting in
cognitive dysfunction, common in FXS and autism. Lauterborn et al. (2007) showed that LTP induced by threshold stimulation was severely impaired in the hippocampal CA1 area in fmr1-KO mice. LTP induced by high frequency stimulation has been studied in different brain regions in fmr1-KO mice. For example, Zhao et al. (2005) reported a complete absence of LTP in the anterior cingulate cortex (ACC) and impaired LTP in the lateral amygdala and proposed that this altered synaptic plasticity was responsible for the aberrant learning patterns and other behavioral phenotypes common to Fragile X mice. Furthermore, Wilson & Cox (2007) found impaired LTP in the visual neocortex of fmr1- KO mice and suggested this may contribute to learning deficits and cognitive dysfunction that is often seen in FXS.
The hippocampus (HC) is crucial for learning and the formation of episodic memory, which has been shown to be impaired in people with ASD (Hare, Mellor & Azmi, 2007). Episodic memory is defined as the conscious recollection of past, personal experiences. As such, the HC must be capable of pattern separation to form distinct representations of
temporal and spatial relationships and pattern completion, because it is unlikely that the same episode is fully replicated. Pattern separation involves sparsification and fanning that are both attributed to the DG (Rolls, 2013), as input from the entorhinal cortex (EC) perforant path is
fanned out over the DG granule cells (GCs), a process that requires high synaptic plasticity. When LTP is impaired at the synapse of the perforant path and the DG GCs of mutant mice due to the absence of NMDA receptors (NMDARs), they show an impaired ability to separate patterns of fear conditioning chambers (McHugh, 2007). Yun and Trommer (2010) found reduced perforant path GC LTP and NMDAR mediated neurotransmission in the DG of fmr1-KO mice. Impaired neurobiological pattern separation in the DG may analogously relate to ASD-like phenotypical impaired pattern processing, like excessive adherence to patterns, impaired ability to detect socially important patterns and impaired episodic memory. Due to its location in the network of information processing, altered neural signaling in the DG is likely to contribute to aberrant learning, memory and other ASD phenotypes with respect to pattern-sensitive information. As such, a focus on synaptic signal transmission may help to further enhance knowledge on the neurobiological mechanisms of ASD and FXS.
The majority of excitatory synaptic transmission is mediated by the ionotropic glutamate receptors NMDA and AMPA. NMDARs are central to producing LTP, whereas AMPA receptors (AMPARs) regulate the maintenance of LTP. Postsynaptic density protein 95 (PSD-95) is a protein that anchors NMDARs and is involved in the stabilization of AMPARs at synapses. Muddashetty, Kelic, Gross, Xu, and Bassell (2007) found that FMRP regulates the local synthesis of PSD-95 and that a dysregulated distribution of AMPARs may alter synaptic transmission, consequently leading to impaired neuronal plasticity in fmr1-KO mice and FXS. The DYRK1A gene encodes certain protein kinases, including the DYRK1A enzyme, that play a crucial role in neuronal development through its regulation of NMDARs (Grau, 2014) and has been implicated in Down Syndrome (Ahn, 2006) and ASD. Mice with a mutated DYRK1A gene were shown to be impaired in the hippocampal dependent Morris water maze memory task. Altered LTP and LTD were also observed, suggesting a role for DYRK1A in synaptic plasticity.
On the other hand, the majority of inhibitory synaptic transmission is mediated by GABAA receptors. El Idrissi et al. (2005) found decreased GABAA receptor expression in seizure-prone fmr1-KO mice, suggesting that the absence of FMRP causes a down regulated GABAergic system, a feature that, as discussed above, may be one of the causal factors of abnormal brain connectivity seen in ASD and FXS. Gephyrin anchors inhibitory
neurotransmitter receptors like GABAA receptors to the postsynaptic cytoskeleton analogous in its function to that of PSD-95 at glutamatergic synapses and has been shown to be down regulated in fmr1-KO mice (Paluszkiewicz, 2011). Down regulation of GABAA receptor functioning is associated with disorders like ASD (Jacob, 2008).
As the DG has not been thoroughly studied in the Fragile X model, a multidisciplinary approach to behavior, electrophysiology and immunohistochemistry of the fmr1-KO mice DG may reveal information about NMDARs, AMPARs and GABARs and provide insight on the role of abnormal functioning of excitatory and inhibitory systems that contribute to the manifestation of autistic behaviors in ASD and FXS with regard to pattern sensitive information.
The current research is expected to provide novel information about DG
electrophysiology in FXS and potentially suggest how this brain region mediates some of the behaviors common to FXS and autism. This will be pursued by conducting the acoustic startle response (ASR) to measure fear, long-term memory and sensorimotor gating, the novel object recognition task (NOR) to measure non-spatial memory and the three-chambered social interaction test (TCSIT) to measure social interaction. Open field (OF) will measure
locomotor activity and anxiety and serve as a manipulation check. In vivo recordings of cortical excitability and field potential recordings from brainslices will be made to measure signal transmission. Finally, immunohistochemistry will be used to analyze the distribution of NMDARs, AMPARs and GABARs. This approach will allow for the integration of data on
behavior, electrophysiology and immunohistochemistry. It is expected that the fmr1-KO mice exhibit aberrant behavior, electrophysiology and receptor distribution.
Methods Animals
Five two-month-old adult male FVB/NJ mice of the FVB/N-129/Ola strain and five two-month-old adult male wild-type (WT) mice will be used. The animals are treated in accordance with the principles and procedures of the National Institutes of Health Guidelines of the Care and Use of Laboratory Animals. This includes maintenance under a controlled temperature of 24 ± 1°C and a humidity of 55 ± 5%) on a 12-hour light (09:00-21:00)/12-hour dark cycle with food and water always available.
Materials
Conditions. The five FVB/NJ mice will form the KO-condition and the three WT mice will form the WT-condition. It was irrelevant to conduct standardization control for gender and age, because all mice are males of the same age.
OF. OF will be used to assess spontaneous activity of the mice and serves as a manipulation check for anxiolytical behavior and locomotion.
NOR. NOR is used to evaluate non-spatial memory function. NOR in rodents is analogous in some ways to human episodic memory (Rajaqopal, Massey, Huang, Oyamada & Meltzer, 2014) . The HC is important for object recognition memory (Broadbent, Gaskin, Squire & Clark, 2010) and the DG for pattern separation during NOR (Bolz, Heigele & Bischofberger, 2015).
ASR. ASR is a sudden, extreme muscular response to a novel or intense auditory stimulus (Olmos-Serrano, Corbin & Burns, 2011) and can provide information about fear and long-term memory. When the stimulus is preceded by a weaker stimulus, the response is
reduced, a process known as pre-pulse inhibition (PPI). ASR and PPI are regulated by the hippocampal DG (Issy, Fonseca, Pardo, Stühmer & Del Bel, 2014) and can be used to measure sensorimotor gating, a process that gates out and integrates sensory information before it is channeled to other brain regions. Reduced sensorimotor gating, possibly due to reduced activation of the GABAergic pathways, is hypothesized to result in hypersensitivity, a feature present in both the phenotype of people with ASD or FXS and fmr1-KO mice. FMRP is required for normal development of the startle response (Yun et al., 2006 ) and PPI is reduced in fmr1-KO mice (de Vrij et al., 2008)
TCSIT. TCSIT is used to evaluate the exploration of a novel social stimulus to assess sociability and interest in social novelty . The human FXS phenotype often includes anxiety, social phobia and avoidance (Cordeiro, Ballinger & Hagerman, 2011). However results of studies on anxiety related phenotypes in fmr1-ko mice are contradictory, ranging from less to more anxiety compared to controls and might be dependent on genetic background. (Kazdoba, Leach, Silverman, Crawley, 2014). Nevertheless, anxiety has been associated to the DG for its role in adult neurogenesis (Wang, David, Monckton, Battaglia & Hen, 2008) and is thus worthwhile investigating in relation to the fmr1-KO DG.
Procedure
Both the KO-condition and the WT-condition will follow the same procedures for behavioral tests, electrophysiology and immunohistochemistry.
OF. Mice will be put in one of the four corners of a slightly covered, opaque arena, with equal lightning across the arena floor, where they will be allowed to explore freely for ten minutes. Locomotion is monitored by horizontal and vertical infrared beams and this information is sent to a computer with ANY-maze video tracking software that automatically assesses the location and activity of the mice so that activity maps can be made. Behavioral output is measured by horizontal and vertical beam breaks. Reduced activity may indicate
motor impairments, but also anxiety if the mice spend more times in the corners of the arena and less time in the open field area. The data will represent the total distance traveled (m), overall average speed (m/s), total time mobile (s), total time immobile (s), total mobile episodes, total immobile episodes and number of line crossings.
NOR. Mice will be put in a slightly covered, opaque test arena individually for 30 minutes before the objects are introduced to the experiment so they can habituate to their environment. The entire experiment will be videotaped to assess subsequent object
exploration. In the training phase two similar Legos objects (F1 and F2), similar in size or smaller than the mice will be introduced. This size will facilitate exploration. The mouse is allowed to explore the objects for five minutes. Contact with the objects includes sniffing, climbing on the object and manipulation and will be coded as well as the time spent with each object and will be recorded with ANY-maze video tracking software. After exploration time the mouse will be returned to its cage only to return the next day for subsequent testing. Once again it will be put in the arena and after 30 minutes one of the familiar objects (F1) and a novel object (N1) that is different in shape and color, will be introduced for exploration for 5 minutes. The amount of time spent with the novel object, relative to the amount of time spent with the familiar object will represent the data and will be calculated as follows: (N1-F1) / (N1+F1). Typically, mice will spend more time exploring the novel object.
ASR. In a sound-proof room, mice from both conditions will be put in a MED-ASR-PRO1 Startle Chamber, individually. During test 1, 65-dB background white noise will be presented continuously to the mice during the experiment and will start 5 minutes before three blocks of acoustic stimuli will be presented, so that the mice can acclimate. During block one, three trials of 40 milliseconds 115-dB startle stimuli with a two milliseconds rise-fall time and an inter-trial interval (ITI) of 30 seconds will be presented, preceded by a 100
two millisecond rise-fall time, preceded by a 100 millisecond null period and 40 milliseconds 65-dB pre-pulse stimuli with a two millisecond rise-fall time, will be presented at an ITI of 30 seconds. Block three will be identical to block one. The next day, during test 2, a similar procedure will take place, with the only difference being that the 75-dB startle stimuli of block two will be replaced by 85-dB startle stimuli. The data will represent flinches measured by a stabilimeter, automatically sent to a computer for analysis. Percent PPI (1-(pre-pulse startle + startle stimulus) / startle stimulus) x 100 will be calculated to measure PPI.
TCSIT. A mouse will be put in the right-hand, closed chamber of a transparent, rectangular box with three chambers and a tiny hole (interaction zone) in the wall that
separates it from the middle chamber (semi-interaction zone), just large enough to stick their snout in, which will allow for interaction with the subject mouse that will be put in the middle chamber with access to the left-hand, opaque chamber (no-interaction zone) and the
interaction zone, for ten minutes. The data represents the time spent with their snout in the interaction zone and time spent in each chamber, as recorded with ANY-maze video tracking software.
Electrophysiology. After the behavioral tests, the 8 mice will be sacrificed and their brain sectioned at 400 µm coronal sections using the Leica tissue chopper. The slices were prepared ice-cold and kept in oxygenated, artificial cerebrospinal fluid before recording. A stimulating bipolar platinum electrode will be placed in the perforant pathway for paired-pulse stimulation (PPS) with inter-stimulus intervals (ISIs) of 50, 100, 150, 200, 250 and 300 milliseconds and recording electrodes will be placed at the Schaffer collateral right after the mossy fiber-CA3 synapse to measure short-term plasticity, expressed as P1 and P2. An AMD machine will be used for stimulation and recording and data will be sampled and digitized at 100kHz using a band-pass filter of 0.3 Hz and 3 kHz. The data will represent percent P2 facilitation or inhibition, calculated as follows: ((P2-P1) / P1) x 100.
Immunohistochemistry. After electrophysiology, two sliced WT and two KO brains will be re-sectioned and prepared for observation with a LEICA SP2 AOBS confocal
microscope using immunohistochemical staining to compare the distribution of NMDARs, AMPARS and GABARs. Of these four brains, 1 WT and 1 KO will be stained with the primary antibodies rabbit polyclonal for PSD-95 from Cell Signaling Technology and mouse monoclonal for DYRK1A from Chemicon to analyze excitatory receptor-associated proteins. The other 1 WT and 1 KO brains will be stained with primary antibodies rabbit polyclonal for Gephyrin from USBiological and mouse monoclonal for the GABAA receptor from Upstate to analyze inhibitory receptor-associated proteins. The secondary antibodies that will be used for immunohistochemical analysis with the confocal microscope are CYP3A goat polyclonal and CYP5A rabbit monoclonal from Life Technologies, where CYP3A binds to the primary antibodies for PSD-95 and Gephyrin and CYP5A binds to the primary antibodies for
DYRK1A and GABAA.
Western blot. Two WT and two KO brains will be dissectioned into CX, HC, CB and diencephalon (DI) and prepared to be loaded in a Biorad western blotting station by diluting them and using the same procedure for immunohistochemical staining that will be used for the brains prepared for observation with the confocal microscope. The secondary antibodies that will be used for western blotting are goat anti-mouse HRP and goat anti-rabbit Igg-HRP from Santa Cruz Biotechnologies, where the anti-mouse binds to the primary antibodies for DYRK1A and GABAA and anti-rabbit binds to the primary antibodies for PSD-95 and Gephyrin.
Data-analysis OF
ANY-maze software will be used to gather the OF data. A MANOVA will be conducted using SPSS 22, because there are seven outcome variables, namely the different
factors of locomotor activity, and there is one categorical predictor variable, namely
condition. It is expected that the KO-condition will score higher on every outcome variable compared to the WT-condition.
NOR
ANY-maze software will be used to gather the NOR data. An independent samples t-test will be conducted, because there is one outcome variable, namely preference for novel object and there is one categorical predictor variable, namely condition. It is expected that the WT-condition spends more time, relatively, exploring the novel object compared to the KO-condition, because of the impaired pattern integration and episodic memory of the fmr1-KO mice.
ASR
ANY-maze software will be used to gather the ASR data. A MANOVA will be conducted, because there are eight outcome variables, namely amount of startle during the different parts of the three blocks and percent PPI. There is one categorical predictor variable, namely condition. It is expected that the KO-condition exhibits a greater startle response during both tests compared to the WT-condition and has a higher percent PPI, because of the impaired PPI.
SIT
ANY-maze software will be used to gather the TCSSNT data. A MANOVA will be conducted, because there are three continuous outcome variable, namely time spent in each zone, and one categorical predictor variable, namely condition. It is expected that the KO-condition spends less time in the full interaction zone compared to the WT-KO-condition. Electrophysiology
A factorial mixed ANOVA will be conducted on the data, because there is one
predictor variables, namely condition and ISI. It is expected that the KO-condition exhibits less percent P2 inhibition compared to the WT-condition or even facilitation in response to perforant path stimulation, expressed as a higher percent P2 facilitation, compared to WT-condition.
Immunohistochemistry
Imaris software will be used to reconstruct the confocal microscope images of the cerebellum (CB), CA3, DG and cortex (CX) and Florochem systems software will be used to analyze the western blot membranes and construct images. Data of the western blot will represent the mean of 4, 2 different exposure modes x bottom/top band. It is expected that the KO-condition shows a higher concentration of PSD-95 and DYRK and a lower concentration of Gephyrin and GABAA receptors compared to the WT-condition.
Results OF
Prior to statistical analysis, no outliers were found and all assumptions were satisfied. A MANOVA revealed that the WT-condition exhibited slightly more locomotor activity than the KO-condition, as shown in Table 1.
Table1
Mean locomotor activity parameters and standard deviations for both conditions.
Locomotor Activity Parameters WT KO
Total distance traveled (m) 71.79 (11.38) 60.66 (11.80) Overall average speed (m/s) 0.12 (0.02) 0.10 (0.02) Total time mobile (s) 539.20 (25.49) 517.40 (26.68)
Total time immobile (s) 60.83 (25.48) 82.60 (26.68) Total mobile episodes 19.00 (6.56) 22.60 (7.02) Total immobile episodes 18.00 (6.56) 21.60 (7.02) Number of line crossings 204.67 (34.49) 157.80 (75.07)
However, no significant main effect of condition on locomotor activity was found, F(1,6)=0,772, p=0.701. Figure 1 shows representative track plots of the OF. These results are not in line with the expectations that the KO-condition would score higher on locomotor activity variables.
Figure 1. Representative track plots for both conditions depicting locomotor activity in the OF.
NOR
Prior to statistical analysis, it was established that all assumptions were satisfied and 2 outliers were identified in the KO-condition. These were not removed from the dataset because of the small sample size. As shown in Figure 2, the preference ratio for the novel object was
considerably lower for the KO-condition compared to the WT-condition.
However, an independent samples t-test revealed no significant difference in
preference for the novel object between the WT-condition (M=0.56, SD=0.45) and the KO-condition (M=0.17, SD=0.41), t (6) = 1.236, p = 0.263. This is due to high standard
deviations, because of the small sample size. The results show correspondence with the expectations that the WT-condition would spend more time exploring the novel object relative to the familiar object. However, because the results were not significant this cannot be
concluded.
Figure 2. Mean preference ratio for the novel object for both conditions.
ASR
After obtaining the results for test 1, differences in mean peak values and percent PPI were observed. Larger values were observed for every variable for the KO-condition during test 1, except for the block 2 pre-pulse startle, as shown in Table 2.
Table2
Mean peak values, percent PPI and standard deviations of startle response of test 1 for both conditions.
WT KO
Block 1 null time 156.67 (32.03) 190.25 (56.41) Block 1 startle 162.33 (76.33) 281.6 (73.68) -0,20 0,00 0,20 0,40 0,60 0,80 1,00 1,20 WT KO P ref er en ce ra ti o Condition
Block 2 null time 119.1 (16.97) 255.72 (39.61) Block 2 pre-pulse startle 288.67 (61.84) 225.36 (59.12) Block 2 startle 127.37 (22.24) 141.1 (33.44) Percent PPI -226.84 (35.79) -160.78 (25.38) Block 3 null time 112.42 (6.01) 282.15 (223.92) Block 3 startle 86.08 (3.26) 191 (34.87)
Prior to statistical analysis of test 1, outliers were found in the KO-condition for block 1 null time, block 1 startle, block 2 null time, block 2 startle and percent PPI. These were, however, not removed because of the small sample size. It was also established that the assumption of normality was violated for block 3 startle, F(1,6) = 13.335, p=0.011. Therefore, an Independent Samples Mann-Whitney U Test was conducted for block 3 startle, which revealed a significant difference between both conditions, p=0.0361. For the other variables, a MANOVA revealed a significant difference between both conditions for block 2 null time F(1,6) = 30.638, p=0.001 and percent PPI, F(1,6) = 9.553, p=0.021. Mean peak values of the startle response for block 2 are shown in Figure 3, where the descending lines indicate a drop in startle response due to PPI.
Figure 3. Pre-pulse inhibition of startle response of test 1 for both conditions.
After obtaining the results for test 2, a difference in mean peak values and percent PPI was observed between the two conditions. Larger values were observed for every variable for the KO-condition during test 2, except for the block 2 pre-pulse startle and the block 3 startle, as shown in Table 3.
Table3
Mean peak values,percent PPI and standard deviations of startle response of test 2 for both conditions.
WT KO Block 1 nulltime 124.63 (19.01) 139.88 (38.87) Block 1 startle 145.23 (33.31) 229.12 (77.04) Block 2 nulltime 137.77 (25.80) 193.62 (54.42) 0 50 100 150 200 250 300 350 400
Nulltime Prepulse Startle
Mea n p ea k v a lu e Parameter WT KO
Block 2 pre-pulse startle 272.83 (136.53) 236.24 (89.47) Block 2 startle 147.73 (6.76) 187.36 (94.85) Percent PPI -187.34 (102.27) -150.29 (76.29) Block 3 nulltime 128.87 (17.03) 159.72 (52.35) Block 3 startle 171.87 (62.72) 149.32 (69.95)
Prior to statistical analysis of test 2, individual outliers were found for block 1 null time and block 2 startle, both in the KO-condition. These were, however, not removed because of the small sample size. All assumptions were satisfied. No significant differences between both conditions were observed. Mean peak values of the startle response for block 2 are shown in Figure 4, where the descending lines indicate a drop in startle response due to PPI. The results of test 1 are in correspondence with the expectations that the KO-condition would exhibit a greater startle response during both tests compared to the WT-condition and a higher percent PPI. The results of test 2 are not fully in correspondence with the expectations.
0 50 100 150 200 250 300 350 400 450
Nulltime pre-pulse Startle
Mea n p ea k v a lu e Parameter WT KO
Figure 4. Pre-pulse inhibition of startle response of test 2 for both conditions. SIT
After obtaining the results, it was observed that the WT-condition spent more time in the interaction zone and less time in the semi-interaction zone and no interaction zone compared to the KO-condition. See Table 4.
Table4
Mean time spent (s) in the different areas of the social interaction box.
WT KO
Interaction zone 268.30 (67.37) 198.00 (28.77) Semi-interaction zone 270.20 (83.87) 306.10 (23.26) No interaction zone 61.40 (31.85) 95.90 (11.15)
Prior to statistical analysis, it was established that there were no outliers and the assumption of normality was violated for time in semi-interaction zone, F(1,6) = 3.543, p=0.023. Therefore, an independent samples Mann-Whitney U Test was conducted for time spent in semi-interaction zone, which revealed no significant difference between both
conditions. Statistical analysis with a MANOVA revealed no significant effects for condition on the other social interaction zones. See also Figure 5. Even though the results are in
correspondence with the expectation that the KO-condition would spend less time in the full interaction zone compared to the WT-condition, they are not significant and thus cannot be confirmed.
Figure 5. Mean time spent in different areas of the social interaction box for both conditions.
Electrophysiology
After obtaining the electrophysiological data, altered paired pulse facilitation was observed for the KO-condition. As shown in Table 6, the mean value of P2 is lower than P1 for every ISI for the WT-condition, indicating paired-pulse inhibition. Less percent P2 inhibition was observed for the KO-condition for every ISI and starting at an ISI of 200, percent P2 facilitation was observed, whereas the WT-condition only showed percent P2 inhibition.
Table6
Mean percent P2 facilitation, inhibition (-) and standard deviations for both conditions.
ISI (msec.) WT KO 50 -60,16 (1.52) -16.67 (6.20) 0 50 100 150 200 250 300 350 400
Interaction zone Semi-interaction zone No interaction zone Ti m e ( s) Zone WT KO
100 -42.51 (2.35) -22.39 (1.47) 150 -29.67 (1.65) -4.78 (2.84) 200 -23.15 (2.33) 4.82 (1.64) 250 -21.22 (4.44) 8.59 (2.11) 300 -17.14 (3.29) 7.72 (2.07)
Prior to statistical analysis, no outliers were found and all assumptions were satisfied. A factorial ANOVA revealed a significant main effect of condition on percent P2
facilitation/inhibition, F(1,47) = 1352.075, p < 0.001, a significant main effect of ISI on percent P2 facilitation/inhibition, F(5,47) = 255.514, p < 0.001, but most importantly a significant interaction effect of condition and ISI was found for the percent P2
facilitation/inhibition, F(5,47) = 18.187, p < 0.001. See Figure 6. These results are in correspondence with the expectations that the KO-condition would exhibit less percent P2 inhibition compared to the WT-condition, or even facilitation.
Figure 6. Mean EPSPs ± SEM expressed as ratio of ((P2-P1)/P1)x100 for both conditions. -70 -60 -50 -40 -30 -20 -10 0 10 20 P er cen ta g e Inter-stimulus interval (ms) WT KO
IHC
Confocal microscope. Analysis of immunohistochemically stained brainslices revealed a difference in the distribution of DYRK, PSD-95, Gephyrin and GABAA receptors between the WT-condition and the condition. As shown in Figures 7-10, the
KO-condition showed a higher density of the proteins PSD-95, DYRK and the colocalization of both in the CB, DG, CA3 and CX compared to the WT-condition.
Figure 7. Distribution of PSD-95, DYRK and colocalization of both in the CB for both conditions.
Figure 8. Distribution of PSD-95, DYRK and colocalization of both in the DG for both conditions.
Figure 9. Distribution of PSD-95, DYRK and colocalization of both in CA3 for both conditions.
Figure 10. Distribution of PSD-95, DYRK and colocalization of both in the CX for both conditions.
As shown in Figures 11-14, the KO-condition showed a lower density of Gephyrin, GABAA receptors and the colocalization of both in the cerebellum, DG, CA3 and cortex compared to the WT-condition. The confocal microscope images confirm the expectation that the KO-condition would show a higher concentration of PSD-95 and DYRK and a lower concentration of Gephyrin and GABAA receptors compared to the WT-condition.
Figure 11. Distribution of Gephyrin, GABAA receptors and colocalization of both in the CB for both conditions.
Figure 12. Distribution of Gephyrin, GABAA receptors and colocalization of both in the DG for both conditions.
Figure 13. Distribution of Gephyrin, GABAA receptors and colocalization of both in CA3 for
Figure 14. Distribution of Gephyrin, GABAA receptors and colocalization of both in the CX for both conditions.
Western blot. Western blot analysis revealed altered expression of DYRK, PSD-95, Gephyrin, GABAA and β-actin for the KO-condition. Figure 15 shows the actual band for DYRK, PSD-95 and β-actin under auto-exposure mode.
Figure 15. Expression of DYRK, PSD-95 and β-actin for both conditions in different brain regions.
When normalized to housekeeping gene β-actin, DYRK and PSD-95 are typically more expressed in the CX, HC and CB of the brain of the WT-condition compared to the KO-condition, whereas expression in the diencephalon is similar. See Figures 16 for DYRK and Figure 17 for PSD-95.
Figure 16. Relative intensity ratios ± SEM of DYRK normalized to β-actin for both conditions. 0 0,2 0,4 0,6 0,8 1 1,2 1,4 1,6 1,8 2 CX HC CB DI R ela tiv e in te n sit y r a tio Brain region WT KO
Figure 17. Relative intensity ratios ± SEM of PSD-95 normalized to β-actin for both conditions.
Figure 18 shows he actual band for Gephyrin, GABAA and β-actin under auto-exposure mode.
Figure 18. Expression of Gephyrin, GABAA and β-actin for both conditions in different brain
regions.
Gephyrin, when normalized to β-actin is typically more expressed in the CX, HC and CB of the brain of the WT-condition compared to the KO-condition, whereas expression in
0 0,5 1 1,5 2 2,5 CX HC CB DI Mea n i n ten si ty ra ti o Brain region WT KO
the diencephalon is similar. See Figure 19. GABAA, when normalized to β-actin, is more expressed in the CX and HC in the brain of the WT-condition compared to the KO-condition, whereas it is more expressed in the CB of the KO-condition and similar in the diencephalon, as shown in Figure 20. These results are contradicting the expectation that the KO-condition would show a higher concentration of PSD-95 and DYRK and a lower concentration of Gephyrin and GABAA receptors compared to the WT-condition.
Figure 19. Relative intensity ratios ± SEM of Gephyrin normalized to β-actin for both conditions. 0 0,2 0,4 0,6 0,8 1 1,2 1,4 1,6 1,8 2 CX HC CB DI Mea n i n ten si ty ra ti o Brain region WT KO
Figure 20. Relative intensity ratios of GABAA ± SEM normalized to β-actin for both conditions.
Discussion
In this study, the functional consequences of altered connectivity in the DG of the Fragile X mouse brain and its relevance to autism were investigated. A multi-disciplinary approach revealed aberrant behavior, synaptic plasticity and excitatory and inhibitory receptor associated proteins and inhibitory receptor distribution for fmr1-KO mice, as hypothesized. How can the altered behavior observed in this study be explained by altered DG
neurobiology?
First of all, because of the DG’s role in pattern separation of information (Leutgeb, 2007), a distortion is likely to contribute to impaired episodic memory, which is analogous to NOR in mice (Rajaqopal et al., 2014). In this study, impaired NOR was observed in fmr1-KO mice. Also, increased excitability in the DG was observed during electrophysiology. This might be due to decreased GABAa receptor expression or due to the increased of excitatory
0 0,2 0,4 0,6 0,8 1 1,2 1,4 1,6 1,8 2 CX HC CB DI m ea n i n ten si ty ra ti o Brain region WT KO
markers found with immunohistochemical analysis of the DG of the fmr1-KO mice. The western blot did not show a significant reduction of GABAa receptor expression in the HC of fmr1-KO mice. However, it should be considered that during western blotting the entire HC is grinded, not just the DG. This might account for the different findings of the western blot and immunohistochemistry. On the other hand, GABAa expression in hippocampal CA3 was also found to be reduced, according to the confocal microscope images. Nonetheless, this indicates that FMRP in the DG is important for NOR and possibly for episodic memory in humans, which is impaired in people with ASD ( HARE).
Secondly, because the DG is involved in the ASR and PPI, alterations in DG signaling found in this study may contribute to altered ASR and PPI. In this study, a reduced PPI and ASR was observed in fmr1-KO mice for test 1 of ASR. Reduced activation of the GABAergic pathways results in hypersensitivity, a phenotype of people with ASD and FXS. The increased excitability observed in the fmr1-KO HC during electrophysiology could be explained by the reduced expression of GABAa receptors. It has been shown by Yun et al. (2006) that FMRP is required for a normal startle response, which was also found in this study. Interestingly, however, no reduction in PPI and less difference between startle response between fmr1-KO mice and control mice in general was observed for ASR test 2, while the amplitude of the startle stimulus was increased. This might be explained by the fact that test 2 was conducted the day after test 1, allowing habituation to occur. This, however, would imply functional memory in fmr1-KO mice.
Thirdly, during the TCSIT fmr1-KO mice were less sociable. Reduced social interest is a key feature of autism (Freitag et al., 2014) and FMRP may thus be involved in this behavioral phenotype. Reduced social interest also be explained by hypersensitivity due to impaired sensorimotor gating involving the DG, so that social stimuli become overwhelming, causing social retraction. Hypersensitivity corresponds with the increased excitability in
fmr1-KO mice that was observed during electrophysiology. Reduced social interest could also be a sign of anxiety and the DG has been implicated in anxiety for its role in neurogenesis (Wang et al., 2008).
However, hyperactive behavior was not observed for fmr1-KO mice during the OF, which contradicts earlier research that has shown fmr1-KO mice to be hyperactive compared to control mice (El Idrissi et al., 2010). During the OF, it was observed that the fmr1-KO mice looked sleepy compared to the control mice. It could have been that the OF was just
conducted at an unsuited moment.
On the other hand, electrophysiology results in this study do confirms hyper excitability in the fmr1-KO mice, as increased excitability was observed in CA1 after stimulating the perforant path. P2 facilitation occurred in fmr1-KO mice at ISIs of 200, 250 and 300. In the control mice no facilitation was found at all, but as the ISI was increased a trend indicated that at some point facilitation would occur in the control mice. However, this facilitation already occurred at an ISI of 200 ms. for the fmr1-KO mice, indicating increased excitability, possibly due to decreased GABAergic signaling and increased excitatory signaling via NMDARs, AMPARS and mGluRs caused by a lack of FMRP. This increased short-term plasticity in the DG is likely to contribute to behavioral phenotypes of fmr1-KO mice and people with ASD and FXS.
Furthermore, confocal microscope images of immunohistochemically stained brainslices made in this study correspond with this hyper excitability in fmr1-KO mice. The images reveal an increase in expression of DYRK1A and PSD-95 in the CX, HC, DG and CB.
However, people with FXS and fmr1-KO mice have elongated, immature synaptic spines. This causes an abundance of NMDARS and AMPARS at these spines, but because they are immature, these receptors could be non-functional. The functionality of these receptors cannot be observed with a confocal microscope, it merely shows the
immunoreactivity of the anti-bodies. The images also revealed a decrease in expression of Gephyrin and GABAA receptors in fmr1-ko mice. This also corresponds with the increased excitability found during electrophysiology. Absence of FMRP might thus be responsible, via increased excitability, for the development of abnormal brain circuitry in ASD and FXS.
Interestingly, the western blot did not reveal this pattern of increased DYRK1A and PSD-95 and reduced Gephyrin and GABAA receptor expression in fmr1-KO. This might be due to the fact that in a western blot analysis, an entire brain region is grinded and used for quantification, whereas with a confocal microscope a tiny part of a brain region is observed.. When observing the western blot graphs, it is also noteworthy that the expression of GABAA receptors is considerably less than that of Gephyrin for both conditions. As Gephyrin anchors GABAA receptors, it would be expected to find about the same amount of Gephyrin as GABAA receptors, but this was not the case. This can be interpreted as that the western blot images also shows immunoreactivity of Gephyrin anti-bodies that have not bonded to any GABAA receptors. Since, we did not stain for NMDARs or AMPARs in particular, this non-specific binding could also have occurred for PSD-95 and DYRK1A.
Next, it should be mentioned that the results should be considered carefully, as the number of mice that participate in this research was small, so the statistical power was low. Future research should increase the number of participants to increase statistical power.
Also, including exploration of synaptic spines in future research might allow for better integration of the confocal microscope images and the western blot, as the factor of immature spines leading to increased expression of excitatory synapse related proteins can be analyzed.
Nonetheless, aberrant behavior, learning and DG electrophysiology were observed in fmr1-KO mice, suggesting that abnormal functioning of excitatory and inhibitory systems due to the absence of FMRP contribute to the manifestation of autistic behaviors in ASD and
FXS. The DG is therefore a candidate for future researchers that try to enhance understanding of the underlying neurobiological mechanisms of ASD and FXS.
Literature
Broadbent, N. J., Gaskin, S., Squire, L. R., & Clark, R. E. (2010). Object recognition memory and the rodent hippocampus. Learning & Memory, 17, 5-11. doi:10.1101/lm.1650110 Bolz, L., Heigele, S., & Bischofberger, J. (2015). Running improves pattern separation during
nobel object recognition. Brain Plasticity, 1, 129-141. doi:10.3233/BPL-150010
Rajaqopal, L., Massey, B. W., Huang, M., Oyamada, Y., & Meltzer, H. Y. (2014). The novel object
recognition test in rodents in relation to cognitive impairment in schizophrenia. Current Pharmaceutical Design, 20, 5104-5114. doi:
10.2174/1381612819666131216114240
Cordeiro, L., Ballinger, E., & Hagerman, R. (2011). Clinical assessment of DSM-IV anxiety disorders in fragile x syndrome: Prevalence and characterization. Journal of
Neurodevelopmental Disorders, 3, 57-67. doi:10.1007/s11689-010-9067-y
Issy, A. C., Fonseca, J. R., Pardo, L. A., Stühmer, W., & Del Bel, E. A. (2014). Hippocampal ether- à-go-go1 potassium channels blockade: Effects in the startle reflex and prepulse inhibition. Neuroscience Letters, 559, 13-17. doi:10.1016/j.neulet.2013.11.026
Olmos-Serrano, J. L., Corbin, J., & Burns, M. P. (2011). The GABAa receptor agonist THIP ameliorates specific behavioral deficits in the mouse model of fragile X syndrome. Developmental Neuroscience, 33, 395-403. doi:10.1159/000332884
Yun, S., Platholi, J., Flaherty, M. S., Fu, W., Kottmann, A. H., & Toth, M. (2006). Fmrp is required for the establishment of the startle response during the critical period of auditory development. Brain Research, 1110,159-165.
De Vrij, F. M. S., Levenga, J., van der Linde, H. C., Koekkoek, S. K., De Zeeuw, C. I., Nelson, D. L.,…Willemsen, R. (2008). Rescue of behavioral phenotype and neuronal protrusion morphology in fmr1-ko mice. Neurobiology of Disease, 31, 127-132. doi:10.1016/j.nbd.2008.04.002
Wang, J., David, D. J., Monckton, J. E., Battaglia, F., & Hen, R. (2008). Chronic fluoxetine stimulates maturation and synaptic plasticity of adult-born hippocampal granule cells. The Journal of Neuroscience, 28, 1374-1384. doi:10.1523/JNEUROSCI.3632-07.2008 Yun, S. H., & Trommer, B. L. (2010). Fragile x mice: Reduced long-term potentiation and n-
methyl-d-aspartate receptor-mediated neurotransmission in dentate gyrus. Willsey, A. J., & State, M. W. (2015). Autism spectrum disorders: From genes to
neurobiology. Current Opinion in Neurobiology, 30, 92-99.
Mandy, W., & Lai, M. (2016). Annual research review: The role of environment in the developmental psychopathology of autism spectrum condition. Journal of Child Psychology and Psychiatry, 57, 271-292. doi:10.1111/jcpp.125014
El Idrissi, A., Ding, X., Scalia, J., Trenkner, E., Brown, W. T., & Dobkin, C. (2005). Decreased GABAa receptor expression in the seizure-prone fragile x mouse. Neuroscience Letters, 377, 141-146. doi:10.1016/j.neulet.2004.11.087
El Idrissi, A., Neuwirth, L. S., & L’Amoreaux, W. (2010). Taurine regulation of short term synaptic plasticity in fragile x mice. Journal of Biomedical Science, 17,15, 1-5. Yang, T., Zhao, H., Lu, Li, X., Xie, Y., Fu, H., et al. (2016). Synaptic plasticity, a prominent
contributor to the anxiety in fragile x syndrome. Neural Plasticity, vol. 2016, 1-12. doi:10.1155/2016/9353929
Ahn, K., Jeong, H. K., Choi, H., Ryoo, S., Kim, Y. J., Goo, J., & et al. (2006). DYRK1A BAC transgenic mice show altered synaptic plasticity with learning and memory defects. Neurobiology of Disease, 22, 463-472. doi:10.1016/j.nbd.2005.12.006
Bear, M. F., Huber, K. M., & Warren, S. T. (2004). The mGluR theory of fragile x mental retardation. Trends Neuroscience, 7, 370-377. doi: 10.1016/j.tins.2004.04.009 Coghlan, S., Horder, J., Inkster, B., Mendez, M. A., Murphy, D. G., & Nutt, D. J. (2012).
GABA system dysfunction in autism and related disorders: From synapse to symptoms. Neuroscience and Biobehavioral Reviews, 36, 2044-2055. doi:10.1016/j.neubiorev.2012.07.005
Cooke, S. F., & Bliss, T. V. P. (2006). Plasticity in the human central nervous system. Brain, 129, 1659-1673. doi:10.1093/brain/awl082
Courchesne, E., & Pierce, K. (2005) Why the frontal cortex in autism might be talking only to itself: Local over-connectivity but long-distance disconnection. Current Opinion in Neurobiology, 15, 225-230. doi: 10.1016/j.conb.2005.03.001
Freitag, C., & Konrad, K. (2014). Autism spectrum disorder: Underlying neurobiology. Journal of Neural Transmission, 121, 1077-1079. doi:10.1007/s00702-014-1270-7
Grau, C., Arató, K., Fernández-Fernández, J. M., Valderrama,A., Sindreu, C., Fillat, C., et al. (2014). Dyrk1a-mediated phosphorylation of glun2a at ser1048 regulates the surface expression of channel activity of glun1/glun2a receptors. Frontiers in Cellular Neuroscience, 8, 331, 1-13. doi:10.3389/fncel.2014.00331
Hare, D. J., Mellor, C., Azmi, S. (2007). Episodic memory in adults with autistic spectrum disorders: Recall for self- versus other- experienced events. Research in
Developmental Disabilities, 28, 317-239. doi:10.1016/j.ridd.2006.03.003
Harris, S. W., Hessl, D., Goodlin-Jones, B., Ferranti, J., Bacalman, S., Barbato, I., et al. (2008). Autism profiles of males with fragile x syndrome. American Journal on Mental Retardation, 113, 427-438. doi: 10.1352/2008.113:427–438
(2012). Trajectories of early brain volume development in fragile x syndrome and autism. Journal of the American Academy of Child and Adolescent Psychiatry, 51, 921-933. doi:10.1016/j.jaac.2012.07.003
Hinton, V. J., Brown, W. T., Wisniewski, K., & Rudelli, R. D. Analysis of neocortex in three males with fragile x syndrome (1991). American Journal of Medical Genetics, 41, 289-294. doi: 10.1002/ajmg.1320410306
Jacob, T. C., Moss, S. J., & Jurd, R. (2008). GABAa receptor trafficking and its role in the dynamic modulation of neuronal inhibition. Nature Reviews Neuroscience, 9, 331-343. Kana, R. K., Uddin, L. Q., Kenet, T., Chugani, D., & Müller, R. (2014). Brain connectivity in
autism. Frontiers in Human Neuroscience, 8, 349, 349, 1-4. doi:10.3389/fnhum.2014.00349
Kazdoba, T. M., Leach, P. T., Silverman, J. L., & Crawley, J. N. (2014). Modeling fragile X syndrome in the Fmr1 knockout mouse. Intracable & Rare Diseases Research, 3, 118-133. doi: 10.5582/irdr.2014.01024
Lauterborn, J. C., Rex, C. S., Kramár, E., Chen, L. Y., Pandyarajan, V., Lynch, G., & et al. (2007). Brain-derived neurotrophic factor rescues synaptic plasticity in a mouse model of fragile x syndrome. The Journal of Neuroscience, 27, 10685-10694.
doi:10.1523/JNEUROSCI.2624-07.200
Leutgeb, J. K., Leutgeb, S., Moser, M. B., & Moser, E. L. (2007). Pattern separation in the dentate gyrus and CA3 of the hippocampus. Science, 315, 961-966. doi:
10.1126/science.1135801
Massey, P. V., & Bashir, Z. I. (2007). Long-term depression: Multiple forms and implications for brain function. Trends Neuroscience, 30, 176-184. doi:10.1016/j.tins.2007.02.005 McDevitt, N., Gallagher, L., & Reilly, R. B. (2015). Autism spectrum disorder (ASD) and
fragile x syndrome (FXS): Two overlapping disorders reviewed through
electroencephalography – What can be interpreted from the available information? Brain Sciences, 5, 92-117. doi:10.3390/brainsci5020092
McFadden, K., & Minshew, N. J. (2013). Evidence for dysregulation of axonal growth and guidance in the etiology of ASD. Frontiers in Human Neuroscience, 7, 671, 1-10. doi: 10.3389/fnhum.2013.0067
McHugh, T. J., et al. (2007). Dentate gyrus NMDA receptors mediate rapid pattern separation in the hippocampal network. Science, 317, 94-99. doi: 10.1126/science.1139158 Muddashetty, R. S., Kelic, S., Gross, C., Xu, M., & Bassell, G. J. (2007). Dysregulated
metabotropic glutamate receptor-dependent translation of AMPA receptor and
postsynaptic density-95 mRNAs at synapses in a mouse model of fragile x syndrome. The Journal of Neuroscience, 27, 5338-5348. doi:10.1523/JNEUROSCI.0937-07.2007 Paluszkiewicz, S. M., Martin, B. S., & Huntsman, M. M. (2011). Fragile x syndrome: The
GABAergic system and circuit dysfunction. Developmental Neuroscience, 33, 349-364.
Pfeiffer, B. E., & Huber, K. M. (2009). The State of Synapses in fragile x syndrome. Neuroscientist, 15, 549-567. doi: 10.1177/1073858409333075
Rolls, E. T. (2013). The mechanisms for pattern completion and pattern separation in the hippocampus. Frontiers in Systems Neuroscience, 7, 74, 1-21.
doi:10.3389/fnsys.2013.00074
Saitoh, O., Karns, C. M., & Courchesne, E. (2001). Development of the hippocampal formation from 2 to 42 years: MRI evidence of smaller area dentate in autism. Brain, 124, 1317-1324.
Methyl-D-Asparte receptor-mediated neurotransmission in dentate gyrus. Journal of Neuroscience Research, 89, 176-182.
Yu, T. W., & Berry-Kravis, E. (2014). Autism and fragile x syndrome. Seminars in Neurology, 34, 258-265. doi:10.1055/s-0034-1386764
Van der Molen, M. J. W., Stam, C. J., & Van der Molen, M. W. (2014). Resting-state EEG oscillatory dynamics in fragile x syndrome: Abnormal functional connectivity and brain network organization. PLoS ONE, 9, (2): e88451
doi:10.1371/journal.pone.0088451
Vissers, M. E., Cohen, M. X., & Geurts, H. M. (2012). Brain connectivity and high functioning autism: A promising path of research that needs refined models, methodological convergence, and stronger behavioral links. Neuroscience and Biobehavioral Reviews, 36, 604-625. doi:10.1016/j.neubiorev.2011.09.003
Wilson, B. M., & Cox, C. L. (2007). Absence of metabotropic glutamate receptor-mediated plasticity in the neocortex of fragile x mice. Proceedings of the National Academy of Sciences, 104, 2454-2459. doi:10.1073 pnas.0610875104
Zhao, M., Toyoda, H., Ko, S. W., Ding, H., Wu, L., & Zhuo, M. (2005). Deficits in trace fear memory and long-term potentiation in a mouse model for fragile x syndrome. The Journal of Neuroscience, 25, 7385-7392. doi: 10.1523/JNEUROSCI.1520-05.2005
