In contrast, in individuals with ASD, one MET risk allele was sufficient to give rise to the atypical pattern of functional activity, showing less deactivation than the nonrisk group. In fact, when comparing those with one risk allele, individuals with ASD exhibited
significantly less deactivation in these regions compared to TD subjects, indicative of an even more atypical phenotype in the clinical population with the GABA activation same MET risk genotype. Consistent with the ROI analysis, a whole-brain comparison of TD versus ASD subgroups within the heterozygous risk group found stronger and more widespread differences than those observed when comparing the TD and ASD groups across genotype ( Figure S1B; Table S3). Based on prior reports of altered DMN function TGF-beta inhibitor in ASD (Kennedy et al., 2006; Kennedy and Courchesne, 2008) and MET’s high expression in the PCC (Judson et al., 2011a), as well as our finding of atypical DMN deactivation in MET risk carriers, we next examined the extent to which the MET functional risk variant modulates intrinsic DMN functional connectivity. We used a seed centered in the PCC ( Fox et al., 2005) for whole-brain functional connectivity analyses in rs-fcMRI
data in a matched sample of 33 TD and 38 children and adolescents diagnosed with ASD. The results were remarkably consistent with the functional activation findings: the MET risk genotype significantly modulated functional connectivity, such that those in the highest risk group (CC; n = 16) had reduced intrinsic connectivity between the PCC and MPFC as well as other nearby regions in the PCC compared to the nonrisk group (n = 16; Figure 2A; Table S4). In agreement with the functional activation analyses, the heterozygous risk group diagnosed with ASD (n = 24) showed a pattern of functional connectivity similar
to that observed in the homozygous risk group, whereas functional connectivity before in the TD heterozygous risk group (n = 15) was no different than the homozygous nonrisk group. Collapsed across genotype, the ASD group exhibited reduced PCC-MPFC connectivity relative to the TD group ( Figure 2B). A whole-brain analysis comparing TD and ASD groups independent of genotype revealed similar, and even more extensive, reductions in DMN connectivity as a function of ASD diagnosis ( Figure S2B). This diagnostic effect appeared to be partially driven by a stronger penetrance of the MET risk allele in the ASD group, as significant differences between TD and ASD subgroups were only observed in risk carriers ( Figure 2B); indeed, MET genotype explained 1.7 times as much variance in DMN connectivity in autistic relative to neurotypical individuals. Using an additional seed within the MPFC, we confirmed that both short- and long-range intrinsic DMN functional connectivity was reduced as a function of both MET risk genotype and ASD diagnosis ( Figure S2D; Table S5).