Impairments in sensory-motor gating and information processing in a mouse model of Ehmt1 haploinsufficiency

Regulators of chromatin dynamics and transcription are increasingly implicated in the aetiology of neurodevelopmental disorders. Haploinsufficiency of EHMT1, encoding a histone methyltransferase, is associated with several neurodevelopmental disorders, including Kleefstra syndrome, developmental delay and autism spectrum disorder. Using a mouse model of Ehmt1 haploinsufficiency (Ehmt1D6Cre/+), we examined a number of brain and behavioural endophenotypes of relevance to neurodevelopmental disorders. Specifically, we show that Ehmt1D6Cre/+ mice have deficits in information processing, evidenced by abnormal sensory-motor gating, a complete absence of object recognition memory, and a reduced magnitude of auditory evoked potentials in both paired-pulse inhibition and mismatch negativity. The electrophysiological experiments show that differences in magnitude response to auditory stimulus were associated with marked reductions in total and evoked beta- and gamma-band oscillatory activity, as well as significant reductions in phase synchronisation. The pattern of electrophysiological deficits in Ehmt1D6Cre/+ matches those seen in control mice following administration of the selective NMDA-R antagonist, ketamine. This, coupled with reduction of Grin1 mRNA expression in Ehmt1D6Cre/+ hippocampus, suggests that Ehmt1 haploinsufficiency may lead to disruption in NMDA-R. Taken together, these data indicate that reduced Ehmt1 dosage during forebrain development leads to abnormal circuitry formation, which in turn results in profound information processing deficits. Such information processing deficits are likely paramount to our understanding of the cognitive and neurological dysfunctions shared across the neurodevelopmental disorders associated with EHMT1 haploinsufficiency.


PCR for deletion-specificity
To establish Dach1-Cre specificity and verify the accuracy of the PCR methods for genotyping, sections from the prefrontal cortex (PFC) and the hippocampus were taken for positive verification of the deleted allele in the mutants (985-bp band in the gel). While a section of the cerebellum, a region where Dach1 is not expressed was taken to check for the non-deleted floxed allele (792-bp). A positive control for the wild type allele was included (95-bp). In addition sections from control PFC, hippocampus and cerebellum animals, which are expected to have one copy of the floxed-allele, were also examined. PCR primers for Ehmt1 Flp/-Forward (5-GCCTGGTGAATTTTAGTGGGC-3); Reverse 1 (5-GTTTGGGGCAAGTGTGGAG-3) Reverse 2 (5-TTGTACAAGAAAGCTGGGTCT-3). The PCR profile used was 94°, 5min, 94°, 40s, 60°, 45s, 72°, 1min for 35 cycles and 72°, 10min.

Electrode implantation
Mice were anesthetized with 2% isoflurane and underwent stereotaxic surgery for the implantation of five enameled stainless steel electrodes (diameter 70 m, California Fine Wire Company): two bilateral frontal electrodes, one monopolar and one bipolar (2.7-mm anterior, 1.5-mm lateral, 1.2-deep relative to bregma); two bilateral hippocampal electrodes, one monopolar and one bipolar (2.7 mm posterior, 3-mm lateral, 2.2-mm deep relative to bregma); and one bipolar electrode in the auditory cortex (2.7-mm posterior, 4-mm lateral, 1.1-deep relative to bregma) (Supplementary Figure 1). The two monopolar electrodes were implanted on the right hemisphere and differential signals between these two electrodes were recorded. Bipolar electrodes were implanted in the left hemisphere. Two epidural screw electrodes were placed above the cerebellum to ground the animal. All electrodes were connected to a Precidip (Switzerland) connector and secured with the use of superglue and dental cement (acrylic and Metabond cement (Sun Medical Co., Japan)). Animals were allowed 7-days recovery before recordings took place. All data reported here are from the differentially recorded signal between the right frontal and hippocampal monopolar electrodes. Previous studies demonstrate that this electrode configuration produces AEP components most characteristically similar to human EEG central cortical scalp recordings (Siegel et al., 2003;Ehrlichman et al., 2008).
Auditory stimuli (50ms sinusoid at 1.5 and 3kHz) were generated using the same software and were delivered through a Power1401 interface (CED). Auditory stimuli were delivered with speakers positioned directly in front of each recording cage. The sound pressure was calibrated at 90dB by using a sound meter based on the approximate height of the animal's head within the plexiglass cage. Mice were given 30mins to acclimate for the first 2 recording trials and 15-mins thereafter. Mice were recorded in the dark, and during the dark cycle in order to record during their normal active state.
Data analysis was performed using MatLab (MathWorks, Natick, MA) software. Due to the loss of EEG signals in later recording sessions, some mice were excluded from the analysis.
ERPs were obtained by averaging epochs centred at Time 0 and 500msec to 0 V, respectively.
The wavelet transform produces a time-frequency representation of power (TF energy).
Measurements can be applied to either the averaged evoke potential or to the individual trials, thus generating either evoked power or total power, respectively. For total power, the noise remains in the signal, which may mask any activity that does not have a high signal-to-noise ratio. Thus for total power a baseline correction is applied. The baseline pre-stimulus period from -600 to -200 was subtracted from the entire peri-stimulus period and applied to the frequency bands independently.
In order to extract the phase-locking factor of the oscillatory burst the normalized complex timevarying power of each trial was averaged (Tallon-Baudry et al., 1996). By averaging, complex values are produced, which includes phase information at each time-frequency region around t and f0. The phase-locking factor is the calculated by unit normalizing or transforming the magnitude information of the complex value. The value remaining represents phase information and is an integer from 0 to 1, which ranges from non-phase-locked or 0, to strictly phase locked or 1. Such methods are found to be robust against artefacts.

Permutation testing
Statistical tests on the frequency data was performed using the permutation method (Westfall and Young, 1993), with the maximal number of iterations allowed by the compared samples or 1,000 iterations in case it surpasses this number. The number of iterations was dependent on the total number of possible permutations (which varied with the number of animals included in analysis).
In the paired-pulse analysis, t-tests were calculated in between controls and mutants around the first pulse. In the MMN analysis, t-tests were calculated in the mutant between 24 th pulse and the deviant pulse; and in the control between 24 th pulse and the deviant pulse. Separate permutation plots were generated for the saline and ketamine condition. In addition, t-tests were calculated between mutant and controls for the ketamine condition and for the saline condition. Statistical permutation plots were generated in all of the above conditions for total power and evoked power. suggesting a delayed N1 or the MMN response. These values did not reach significance, due to level of variability. After ketamine administration we find the classic N1 time epoch 25-50ms as well as 50-75ms were significantly different from zeros. However, these values were below zero, thus the standard condition wavelets were larger than deviant condition suggesting the predicted loss of MMN detection after ketamine administration occurred in controls. B) MMN area wavelet differences across 8 time epochs for the Ehmt1 D6cre/+ mutant mice after administration of saline (triangles) and ketamine (circles).

Ehmt1
There was a slightly above 0 value in mutant mice at the 75-100 epoch, however overall at no epoch was there a significant difference from 0 and little to no change in the overall pattern of response was seen between the ketamine and saline condition in these mice. Data are means ± SEM; $ represents significant difference in ketamine condition (p<0.05; $$, p<0.01)