The Dup15q Syndrome EEG Signature Resembles the EEG Effect of GABA-a Modulators

Background:

Duplication 15q11.2-q13.1 (Dup15q) syndrome is a neurogenetic syndrome that is highly penetrant for autism spectrum disorder (ASD), as well as intellectual disability (ID) and epilepsy. Excess beta (12-30 Hz) oscillations in the electroencephalogram (EEG) are the distinguishing electrophysiological feature of Dup15q syndrome that may relate to clinical traits [Frohlich et al, 2016]. It is uncertain to what extent this EEG signature relates to expression of gamma-aminobutyric acid-A receptor (GABAaR) genes GABRA5, GABRB3, GABRG3 in the duplicated 15q11.2-q13.1 locus.

Objectives: We sought to compare the Dup15q syndrome power spectrum to spectral changes induced by the benzodiazepine midazolam in healthy volunteers (HV). Midazolam and other benzodiazepines are non-selective GABAaR positive allosteric modulators (PAMs) that enhance the inhibitory chloride current through the GABAaR in the presence of GABA. In this manner, midazolam may provide an acute reference model for the electrophysiological effects of GABAaR gene overexpression as presumed in Dup15q syndrome.

Methods: Spontaneous EEG recordings were obtained from n = 24 children with Dup15q syndrome with ages spanning from 16 months to 14 years. These recordings were directly compared to those from 1) n = 14 age-matched typically developing (TD) children, 2) n = 14 age-and-IQ-matched children with non-syndromic ASD. Power spectral densities (PSDs, 1-32 Hz) were obtained using a Morlet wavelet transform and normalized to give relative power. The analysis was unbiased with respect to directionality, frequency band, and electrode location and accounted for multiple comparisons using cluster randomization statistics. Group differences in spectral power were compared to changes induced by midazolam in n = 12 healthy adult volunteers 1 hour following a 5mg oral administration relative to pre-dose baseline.

Results: Power differences between children with Dup15q syndrome and control groups strongly resembled spectral changes induced by midazolam in HV. Specifically, the Dup15q cohort exhibited a greater than 100% increase in relative beta power mirroring a 20-40% increase induced by midazolam in HV. Statistically significant (p < 0.05) clusters encompassing all channels revealed decreased power peaking at 9.5 Hz and increased power peaking at 22.6 Hz in Dup15q syndrome relative to pooled controls. A subset of electrodes belonged to a cluster of decreased power peaking at 3.8 Hz in Dup15q syndrome.

Conclusions: Spectral power anomalies identified in Dup15q syndrome are consistent with a pattern of increased high frequency oscillations and decreased low frequency oscillations reported in healthy adult volunteers challenged with midazolam and other benzodiazepine compounds. While some caution is required comparing results from adults to children, we nonetheless suggest an electrophysiological isomorphism between the Dup15q syndrome EEG biomarker and midazolam-induced EEG changes. Considering that children with Dup15q syndrome have 1) duplications of GABAaR subunit genes and 2) EEG changes mirroring those induced in HV by the GABAaR PAM midazolam, we conclude that EEG anomalies in Dup15q syndrome likely relate to overexpression of GABAaR genes. In fact, previous work [Urraca et al, 2013] has reported such EEG anomalies in patients with paternal 15q11.2-q13.1 duplications, thus obviating the paternally silenced gene UBE3A as the cause of such EEG patterns.