ASD Risk Polymorphism in MET is Associated with an Aberrant Pattern of Functional Activity Across Regions of High MET Expression

Background: A common variant (rs1858830) in the promoter region of MET receptor tyrosine kinase (MET) has been associated with ASD risk across multiple independent samples (Campbell 2006, Campbell 2008, Jackson 2009). In the primate, MET is enriched in neurons and their axons which project from subcortical limbic forebrain structures -- as well as temporal, parieto-occipital and limited frontal cortices -- to target regions during synaptogenesis (Judson 2010). MET transcript and protein expression in postmortem temporal cortex is reduced in ASD, as well as in neurotypical individuals who carry this common risk variant (Campbell 2007). Interestingly, many regions in which MET is highly expressed are involved in socio-emotional processing and have been found to exhibit abnormal functional activity in ASD (see Di Martino 2009).

Objectives: Despite clear evidence that the MET promoter variant disrupts MET expression, which is crucial for typical circuit development, no studies have examined the effect of this variant in humans using in vivo neuroimaging. Therefore, we sought to examine the effects of the MET risk polymorphism in a relatively large sample of neurotypical subjects and individuals with ASD using functional magnetic resonance imaging (fMRI) during an emotional face processing task (Dapretto 2006, Pfeifer 2008).

Methods: While undergoing fMRI, 23 children and adolescents with ASD and 59 typically-developing subjects (matched by age, IQ and head motion) passively observed faces displaying different emotions (angry, fearful, happy, sad, and neutral). Using a jittered event-related design, each facial expression was presented 16 times for 2 seconds. DNA was extracted from saliva samples and rs1858830 was genotyped using a custom Taqman assay. Random-effects group-level analyses were performed to examine the main effects of ASD diagnosis and MET risk polymorphisms on emotional facial processing (genotype groups were matched by ASD diagnosis, age, IQ and head motion).

Results: Regardless of diagnosis, analyses of individuals possessing one or two MET risk alleles showed hyperactivated subcortical structures that receive MET cortical projections, including thalamus, striatum, and amygdala, as well as cortical regions including inferior frontal gyrus and supplementary motor cortex, according to an additive genetic model. Additionally, risk carriers displayed reduced deactivation of MET-enriched cortical regions, particulary temporo-parietal regions that include primary auditory cortex, planum temporale, superior temporal sulcus, and parietal operculum. When examining the main effect of ASD diagnosis, a similar pattern emerged whereby nearly all of the regions displaying greater activity in risk carriers displayed greater activity in children with ASD (most significantly in temporal regions). Preliminary analyses did not find evidence of an interaction between MET risk alleles carrier and ASD diagnosis.

Conclusions: While autism is a highly heritable neurodevelopmental disorder, relatively little is known about how genetic factors alter the development of circuitry underlying social and emotional processing. Our findings suggest a relationship between a risk allele that reduces MET expression and disrupted activity in specific subcortical and cortical circuits where MET is highly expressed. These findings lend support to the notion of a broader autism phenotype and highlight how an imaging genetics approach may further elucidate the neurobiology of ASD.