Left Parietal Functional Connectivity Mediates the Association Between COMT rs4633 and Verbal Intelligence in Healthy Adults

In Chinese Han population, Catechol-O-methyltransferase gene (COMT) rs4633 is found to be associated with impaired cognitive process. We aimed to investigate the association between COMT rs4633 and verbal intelligence and the underlying neural mechanisms in Chinese Han healthy young adults. In 256 Chinese Han healthy young adults, we explored the modulatory effects of COMT rs4633 on verbal intelligence quotient (VIQ) and functional connectivity density (FCD) of the brain and the mediation effect of FCD on the association between COMT and VIQ. We further investigated the association between the expression patterns of dopamine receptor genes and the effect of COMT on FCD in the human brain. COMT rs4633 TT homozygotes exhibited lower VIQ than CC homozygotes and TC heterozygotes, higher long-range FCD (lrFCD) than CC homozygotes and TC heterozygotes in the left superior frontal gyrus. TT homozygotes and TC heterozygotes showed higher lrFCD than CC homozygotes in the left inferior parietal lobule. The lrFCD differences across genotypic subgroups were negatively associated with the expression of DRD2 and DRD3 genes. The left parietal lrFCD mediated the association between COMT rs4633 and VIQ. These findings provide a biological pathway that COMT rs4633 affects verbal intelligence via modulating the lrFCD of the left inferior parietal lobule.

Three probes were developed for the LDR reactions, including a common probe (rs4633_modify: P-TGGTTCAGGATGCGCTGCTCCTTGGTTTTTTTTTTTTTTTTTTTT-FAM) and two discriminating probes (rs4633_C: TTTTTTTTTTTTTTTTTTTTCCCGGGCTCCGCATGCTGCAGC, rs4633_T: TTTTTTTTTTTTTTTTTTTTTTCCCGGGCTCCGCATGCTGCAGCACA ACG) for two alleles of COMT rs4633. These reactions were carried out in a 10 μL mixture containing 1 μL buffer, 1 μL probe mix, 0.05 μL Taq DNA ligase, 1 μL PCR product, and 6.95 μL deionized water. The reaction program comprised an initial heating at 95 °C for 2 min followed by 35 cycles of 30 s at 94 °C and 2 min at 50 °C. We ceased the reactions by chilling the tubes in an ethanol-dry ice bath and adding 0.5 mL of 0.5 mM EDTA. Aliquots of 1 μL of the reaction products were blended with 1 μL of loading buffer (83% formamide, 8.3 mM EDTA and 0.17% blue dextran) and 1 μL ABI GS-500 Rox-Fluorescent molecular weight marker, denatured at 95 °C for 2 mins, and chilled rapidly on ice prior to being loaded on a 5 M urea-5% polyacrylamide gel and electrophoresed on an ABI 3100 DNA sequencer at 3000 V. Finally, we used the ABI Gene Mapper software to assay and quantify fluorescent ligation products.

fMRI data preprocessing
The resting-state fMRI data were preprocessed using the SPM12 (www.fil.ion.ucl.ac.uk/spm) with the following steps. The first 10 volumes were discarded for signal reaching equilibrium and participants adapting to the scanning noise. Acquisition time delays were corrected between slices for remaining 170 volumes and these volumes were realigned to the first volume to correct head motions. During the correction of head motion, we only included subjects with translational or rotational motion parameters lower than 2 mm or 2°. What's more, we calculated the frame-wise displacement (FD), an index representing volume-to-volume changes in head position (Power et al., 2012). We obtained FD from the derivatives of the rigid body realignment estimates that were used to realign functional MRI data (Power et al., 2012;2013). These functional images were spatially normalized to Montreal Neurological Institute (MNI) space using steps described below: we first linearly co-registered individual structural images into the mean motion-corrected functional image; then we segmented the co-registered structural images into gray matter, white matter and CSF, meanwhile nonlinearly co-registered gray matter to the MNI space; and finally we normalized the motion-corrected functional volumes into the MNI space by using the parameters estimated during nonlinear co-registration. Then, we resampled the normalized functional volumes into a voxel size of 3 × 3 × 3 mm 3 . At last, functional images were band-pass filtered with a frequency range of 0.01-0.1 Hz and nuisance covariates (including six head motion parameters and average BOLD signals of the ventricular, and white matter) were regressed out.

FCD calculation
To minimize unwanted effects from susceptibility-related signal-loss artifacts, a gray matter (GM) mask was applied to restrict the calculation of the FCD to voxels only in the GM regions with a signal-to-noise >50 (Tomasi and Volkow, 2010). To enhance the normality of the distribution, grand mean scaling of FCD was obtained by dividing by the mean FCD value of the qualified voxels in the whole brain. Finally, the normalized FCD maps were spatially smoothed with a 6 × 6 × 6 mm 3 full-width at half maximum (FWHM) Gaussian kernel.

Gene expression analysis
There were four brains only having left-hemispheres and two brains having two hemispheres. Each donated brain had MRI scans before being dissected, so each sample had MNI coordinate. And importantly, mRNA microarray analysis was done for each sample to depict the transcriptional profiles in human brain (detailed information please see http://help.brain-map.org/display/humanbrain/Documentation).
The steps actually performed in this analysis are as follows. Firstly, To get the DRD2 and DRD3 expression values in each sample of six donated brains, we chose one probe to represent gene expression of DRD2 and DRD3 by calculating which probe was most correlated with other probes as done in previous studies (Hawrylycz et al., 2015;Forest et al., 2017), which could make the results more reliable. Secondly, we extracted the average F values indicating group differences of FCD within spherical ROIs (radius = 6mm) centered at MNI coordinate of each sample in each brain. Thirdly, we performed sample-wise spatial correlation between gene expression profiles and effect of COMT on brain FCD for DRD2 and DRD3. Lastly, we did one sample T test for correlation coefficient to compare the consistency among six donated brains. Because we performed two spatial correlation analyses, Bonferroni method was applied to correct for multiple comparisons (P < 0.05).

Demographic information in 279 subjects
After excluding 44 subjects with genotyping failure (n=29) and without cognitive data (n=15), 279 healthy young Chinese Han subjects were included in the cognitive analysis. The genotypic distributions of COMT rs4633 was in Hardy-Weinberg equilibrium (P = 0.431). There was significant difference in years of education (P = 0.009), but not in gender and age (P > 0.05) among the three genotypic subgroups (Table S1).