Benzodiazepine Binding Site and GABA-B Receptor Density in the Cerebellar Cortex, Broca's, and Wernicke's Areas in Individuals with Autism

Background: There are widespread abnormalities across brain areas in multiple GABA receptor subtypes in individuals with autism. Two particular types, benzodiazepine (BZ) binding GABAA receptors (GABA-AR) and GABAB receptors (GABA-BR), are reduced in number in the anterior and posterior cingulate cortex and fusiform gyrus. The Crus II region of the cerebellum in autism has the most pronounced decrease in Purkinje cells (PCs) as well as altered GAD65/67 mRNA in PCs and basket cells. There is a paucity of information regarding GABAergic changes in the speech and language areas in autism.

Objectives: To determine whether there are alterations in density of these receptor subtypes in two different regions of the cerebellar cortex: the vermis (centered on lobule VI) and the lateral hemisphere (Crus II), as well as in two cerebral cortical areas for speech and language: Broca's area (BA 44/45) and Wernicke's area (BA 22).

Methods: Age-matched postmortem brain tissue from Crus II (n=17 autism; n=21 control), vermis lobule VI (n=7-9 autism; n=11 control), Broca's area (n=14-15 autism; n=17 controls) and Wernicke's area (n=10-12 autism; n=12-13 control) were processed for ligand binding autoradiography for ³[H]-flunitrazepam (2 nM) labeled GABA-AR and ³[H]-CGP54626 (1.5nM) labeled GABA-BR. Binding density was quantified in the molecular and granular layers in the cerebellum or in the superficial (I-IV) and deep (V-VI) lamina in the cerebral cortical areas.

Results: Significant decreases in the density of BZ binding sites were demonstrated in both the molecular (p=0.0007; two tailed unpaired t-test) and granular layers (p=0.0034) in the Crus II region in the autism group. In contrast, normal density of BZ binding sites was found in vermis lobule VI.  Significant reductions in BZ sites were also found in Broca's (p=0.0162 superficial layers; p=0.0004 deep layers) and in Wernicke's areas (p=0.0008 deep layers) but not in superficial layers.  All four brain areas contained normal density of GABA-BRs.

Conclusions: Within the cerebellum there are disparate findings indicating that the Crus II region, which receives pronounced frontal cortical inputs via the pons and is reported to contain reduced numbers of PCs in individuals with autism, has markedly reduced density of BZ binding sites on GABA-ARs throughout its layers.  In contrast, the vermis, which receives spinocerebellar and other motor inputs, contains normal BZ site density. These findings suggest a selective inhibitory dysfunction in the lateral hemisphere that may affect modulation of frontal cortical-related activity. The decrease in BZ density in both language areas confirms the pervasive GABA-AR disturbances across brain areas in autism. On the other hand, GABA-BR abnormalities are so far limited to specific cerebral cortical areas and do not include Broca's and Wernicke's areas or the cerebellar cortex.