GABA(A) Receptors in Autism - a Multicenter Positron Emission Tomography Study

Background: GABA (gamma-amino-butyric-acid) is the primary inhibitory neurotransmitter in the human brain. It has been proposed, on the basis of animal, genetic and post-mortem evidence, that the symptoms of autism spectrum disorders (ASDs) are associated with deficient GABA neurotransmission, possibly including reduced expression of GABA(A) receptors. These receptors are also a major focus of drug discovery (and recent Fast Fail initiatives in the USA). We recently (Mendez et al. 2013) reported preliminary evidence of reduced GABA(A) alpha5 binding in vivo in adults with an ASD. However, this requires replication. Moreover it is unknown whether any putative differences in GABA(A) are generalised to all receptor subtypes, or specific to alpha5.

Objectives: We conducted the first large study to investigate GABA(A) receptor density in adults with 'high functioning' ASD, using Positron Emission Tomography (PET). In this multicenter study, part of the EU-AIMS consortium, we carried out two linked experiments  to provide complementary information about GABA(A). In one, we used the tracer [11C]flumazenil for quantification of total GABA(A) receptors; and in the other, we used [11C]Ro154513 as this tracer allows the selective quantification of GABA(A) alpha1 and alpha5 subtypes. Together, these two tracers provide a comprehensive analysis of GABA(A) density..

Methods: In the [11C]flumazenil experiment, 15 adults with ASD were compared to a control group of 8 healthy volunteers matched for age, sex, and IQ. Regional binding potential (BP_ND) to the GABA(A) receptor was examined. Quantitative analysis of PET images used a simplified reference tissue model with the pons as reference region, as standard for this tracer (Klumpers et al. 2008).

In the [11C]Ro154513 experiment, 10 adults with ASD were compared to 7 healthy controls matched for age, sex, and IQ. [11C]Ro154513 V_T was calculated on the basis of an arterial input function and GABA(A) alpha1 and alpha5 receptor binding was estimated using a spectral decomposition of the activity time course (Myers et al. 2012).

Across both studies, inclusion criteria included: age 18-50; medically fit; free of psychotropic medication at the time of the study; IQ > 70; no history of epilepsy or other neurological illness. All ASD volunteers were diagnosed by using DSM-V criteria and verified by ADOS and, where possible, ADI. None of the volunteers in the [11C]flumazenil experiment took part in the [11C]Ro154513 experiment.

Results: In the [11C]flumazenil experiment, we found no evidence of differences in binding potential in any brain region (all p>0.22). Similarly using [11C]Ro154513, we found no significant differences in whole-brain estimates of either GABA(A) alpha1 (p=0.44) or alpha5 (p=0.49) subtype receptors in adults with ASD, nor were there significant regional differences in any brain area.

Conclusions: This is the largest PET study of GABA(A) in ASD to date. Nonetheless our initial results are inconsistent with the hypothesis that GABA(A) receptor abnormalities are present across the ASD spectrum in adults with normal intellectual ability. However, reductions in GABA(A) receptors may exist in children (Mori et al. 2011), or in other ASD subgroups. We also cannot rule out differences in GABA(A) receptor function.