Sex‐specific effects of COMT Val158Met polymorphism on corpus callosum structure: A whole‐brain diffusion‐weighted imaging study

Abstract Background Genetic polymorphisms play a significant role in determining brain morphology, including white matter structure and may thus influence the development of brain functions. The main objective of this study was to examine the effect of Val158Met (rs4680) polymorphism of Catechol‐O‐Methyltransferase (COMT) gene on white matter connectivity in healthy adults. Methods We used a whole‐brain diffusion‐weighted imaging method with Tract‐Based Spatial Statistics (TBSS) analysis to examine white matter structural integrity in intrinsic brain networks on a sample of healthy subjects (N = 82). Results Results revealed a sex‐specific effect of COMT on corpus callosum (CC): in males only, Val homozygotes had significantly higher fractional anisotropy (FA) compared to Met‐carriers. Volume‐of‐interest analysis showed a genotype by sex interaction on FA in genu and rostral midbody of CC, whereby Val males demonstrated higher FA than Met females. Conclusions These results demonstrate the key effect of genes by sex interaction, rather than their individual contribution, on the corpus callosum anatomy.

of the 22nd chromosome (Bertocci et al., 1991;Deutch & Roth, 1990;Grossman, Emanuel, & Budarf, 1992). The COMT gene is involved in the metabolic degradation of catecholamines, notably synaptic dopamine, especially in the prefrontal cortex (Männistö & Kaakkola, 1999;Mattay et al., 2003;Tunbridge, Bannerman, Sharp, & Harrison, 2004). This gene is associated with several allelic variants, including the COMT Val158Met (rs4680), a functional single-nucleotide polymorphism that codes for a substitution of methionine (Met) for valine (Val) at codon 158. Previous studies have shown that this Met substitution reduces the activity of COMT enzyme to one-quarter of what is originally encoded by the Val allele (Lachman et al., 1996;Lotta et al., 1995).
The COMT Val158Met polymorphism may impact prefrontally mediated cognition (Meyer-Lindenberg & Weinberger, 2006). Several studies showed that Met-carriers performed better in cognitive tasks that involve prefrontal cortex functionality notably working memory and cognitive flexibility tasks, compared with Val homozygotes.
This effect was found independently of diagnosis, for example, in healthy controls, individuals with schizophrenia and schizotypal disorder Malhotra et al., 2002;Minzenberg et al., 2006). However, a meta-analysis (Barnett, Scoriels, & Munafò, 2008) highlighted that the relationship between COMT and cognitive performance was not always evident. More recently, a large populationbased study (Wardle, de Wit, Penton-Voak, Lewis, & Munafò, 2013) has not found any effect of COMT on working memory. The authors suggested that the COMT effects are better detectable in neurophysiological experiments rather than in purely cognitive tasks. fMRI studies have indeed demonstrated COMT effect on Blood Oxygenation Level Dependent (BOLD) signal in people performing working memory tasks (Egan et al., 2001;Mattay et al., 2003). However, the metaanalytical data on this topic has been inconsistent. Meta-analysis by Mier, Kirsch, and Meyer-Lindenberg (2010) reported significant association between the COMT genotype and prefrontal activation in executive function tasks, whereas the most recent meta-analytical paper indicated that there is presently no (significant) spatial convergence of imaging genetics findings on this association (Nickl-Jockschat, Janouschek, Eickhoff, & Eickhoff, 2015).
As well as studying the effect of COMT on brain functions, the investigators have been increasingly looking at possible associations between COMT polymorphisms and brain structure, for example, brain white matter as indexed by fractional anisotropy (FA). Shimony et al. (1999) used quantitative diffusion anisotropy MR images obtained from 13 healthy adults to evaluate white matter integrity.
Their results showed that the FA measure reflects density and myelination of fibers, as well as directional coherence, which could contribute to efficient signal transmission. Thomason et al. (2010) examined COMT gene-related differences in four white matter fiber tracts (genu of the corpus callosum, anterior thalamic radiation, inferior fronto-occipital fasciculus, uncinated fasciculus) in a group of 40 children (aged 9-15). The authors reported a significantly higher FA in Val-allele homozygotes (vs. Met/Val heterozygotes and Met allele homozygotes) in the genu of corpus callosum and anterior thalamic radiation, respectively. Based on animal studies showing increase in DA availability due to inhibition of COMT functionality (e.g., Tunbridge et al., 2004), the authors suggested that lower FA measures in Met-carriers were associated with higher dopamine availability that may have inhibited myelination. Their results were in accordance with those of Liu et al. (2010) who employed a haplotype-based approach to assess the effect of COMT gene polymorphisms on white matter integrity. The authors found that the groups with lower COMT enzymatic activity (i.e., potentially higher dopamine levels) had lower FA values in the prefrontal white matter tracts.
However, the main effect of COMT was not replicated in a study of Kohannim et al. (2012). This voxelwise analysis study reported a combined effect of five genetic SNPs, that is, COMT Val158Met, clusterin CLU, neuregulin 1 receptor (ErbB4), neurotrophic tyrosine kinase receptor-type 1 (NTRK1), and the hemochromatosis gene (HFE) which The studies showed a sex effect on COMT functioning. In particular, COMT activity in human prefrontal cortex is 17% higher in men than women (Chen et al., 2004). The sexual dimorphism in the normal brain could be due to interaction of estrogens with COMT activity. Estrogens inhibit COMT gene transcription (Xie, Ho, & Ramsden, 1999), while COMT, in turn, participates in the metabolism of estrogens (Worda et al., 2003). This sexual dimorphism is known to contribute to significant sex by genotype interaction effects on behavior and personality (Barnett et al., 2008;Chen et al., 2011;Gurvich & Rossell, 2015;Harrison & Tunbridge, 2008) as well as cortical thickness (Papaleo, Erickson, Liu, Chen, & Weinberger, 2012;Sannino et al., 2015). In particular, we expected to observe sexual dimorphism effects, that is, reduced integrity of white matter tracts in females with Val polymorphism (compared with their male counterparts) due to lower activity of Val polymorphism and putatively higher DA availability in females. In order to make our study comparable with the studies reviewed as above, we chose to group COMT genotypes according to the presence or absence of the Met allele (i.e., Met/Met + Val/Met

vs. Val/Val).
We used a whole-brain diffusion-weighted imaging (DWI) tractography method with Tract-Based Spatial Statistics (TBSS) analysis to examine influence of COMT polymorphism on the intrinsic brain networks in a sample of healthy adults. We chose FA as the measure of white matter structure, as it has been shown to have higher heritability than other DWI parameters, such as radial and axial diffusivity (Kochunov et al., 2010).
The White Caucasian ancestry has been established by participants' self-report. Exclusion criteria comprised the neuropsychiatric disorders in the participants or their first-degree relatives, screened out by the Structured Clinical Interview for DSM-IV (First, Spitzer, Gibbon, & Willimans, 1996). Participants were provided with full information about the study and they gave written consent according to declaration of Helsinki. The study was approved by Ethical Committee of King's College London, Institute of Psychiatry (London, UK).

| Genotyping
We acknowledge the potential pitfalls associated with neuroimaging genetic studies looking at the effects of a candidate genes on brain function or structure (e.g., Bigos & Weinberger, 2010). These authors suggested a set of basic principles that would assure sufficient power and validity of such a study, for example, a rational approach to the selection of candidate genes, the careful control of nongenetic factors (i.e., sex, age, IQ), and selection of task paradigms that could be plausibly linked to the biology of the gene of interest. In our investigation of white matter connectivity, we chose a COMT gene that has been reported to be related to white matter integrity. We also modeled effect of sex as potential factor interacting with the effect of COMT polymorphism.
Cheek swab samples were collected and DNA extracted using the protocol developed at the Social, Genetic and Developmental Psychiatry Centre (King's College London, Institute of Psychiatry, London, UK) (Freeman et al., 2003). All participants were genotyped for COMT (Val158Met, rs4680). The genotype of the COMT Val158Met (rs4680) SNP was determined by allelic discrimination assay (C__25746809_50) based on fluorogenic 5′ nuclease activity: a TaqMan SNP genotyping assay was performed using an ABI Prism 7900HT and analyzed with Sequence Detection System software according to the manufacturer's instructions (Applied Biosystems, Warrington, UK).
There was no statistical deviation from Hardy-Weinberg equilibrium for all the polymorphisms. The COMT Val/Val homozygosity was detected in 19 participants (12 males and 7 females) and 63 participants were Met-carriers (27 males and 36 females).

| DWI image acquisition
We acquired DWI that uses the water diffusion properties to investigate the microstructural architecture of white matter fibers (Basser, Mattiello, & LeBihan, 1994). FA is a parameter derived from DWI that provides a reliable surrogate of the white fiber microstructural integrity. All experiments were performed on a 3T scanner (Signa HDx, General Electric, Milwaukee, WI, USA). DWI was acquired on whole-brain volumes with a single-shot echo planar imaging sequence and an 8-channel head coil. Imaging parameters were as follow: TE/ TR = 104/17,910 ms, pixel size = 2.4 × 2.4, slice thickness = 2.4, 32 directions with b value = 1,300 mm 2 /s including four images with no diffusion weighting (b value = 0).

| DWI data preprocessing
Processing of diffusion weighted images was performed using FSL FMRIB's Software Library (FSL, version 4.1.9, http://www.fmrib.ox.ac. uk/fsl) (Smith et al., 2004). Diffusion-weighted data were corrected for eddy-current distortions and simple head motion artifacts using affine registration to the first acquired volume. Then a binary brain mask image was computed for each subject using Brain Extraction Tool (BET) (Smith, 2002). Finally, FA maps were created from the diffusion tensor model employing the FDT diffusion tool.

| Whole-brain TBSS analysis
Voxelwise statistical analysis of the FA data was carried out using TBSS (Tract-Based Spatial Statistics). TBSS is a useful tool for analyzing FA data as it improves sensitivity, objectivity, and interpretability through nonlinear registration with an alignment-invariant tract representation (Smith et al., 2006). All subjects' FA data were aligned into a common space using the FNIRT (FMRIB's Non-Linear Image Registration Tool) (Andersson, Jenkinson, & Smith, 2007), which uses a b-spline representation of the registration warp field (Rueckert et al., 1999). Next, a mean FA image was created and thinned to create a mean FA skeleton which represents the centers of all tracts common to the group. Each subject's aligned FA data were then projected onto this skeleton and the resulting data fed into voxelwise cross-subject statistics. A threshold of FA > 0.2 was applied to exclude areas of uncertain diffusion orientation and/or high intersubject variability.
A permutation program (FSL's Randomize) was applied to voxelwise statistics on skeletonized FA data for all subjects. A design and a contrast matrix were performed using General Linear Model (GLM Tools, part of FSL). Next, randomization procedure produced statistical maps so that each pixel value refers to the p-value of the test. The threshold-free cluster enhancement (TFCE) (Smith & Nichols, 2009) and 5,000 permutations options were applied.
The GLM analysis was performed where we have modeled main effects of the COMT (Met-carriers vs. Val homozygotes) and sex, as well as interactions of COMT x sex. The age was considered as a covariate.

| Volume-of-interest analysis
Although TBSS has been very popular approach for analyzing DWI data, the investigators also reported potential inaccuracies that are inherent in the FA skeleton projection and the substantial bias that it can introduce (see review, Bach et al., 2014). Therefore, in order to maximize the robustness and validity of our results we reevaluated the TBSS data by applying the volume-of-interest analy-

sis (VOI).
According to Witelson's method, six parts of the corpus callosum were examined: rostrum, genu and rostral body, anterior midbody, posterior midbody, isthmus, and splenium (Witelson, 1989). By applying a nonlinear spatial normalization (computed during TBBS procedure), the VOI was co-registered to the FA of each subject. Then, the mean FA was computed in the VOI for each subject. Finally, a statistical analysis was performed for each tract using Kruskal-Wallis test.

| Whole-brain TBSS analysis
The GLM has not produced significant main effects of COMT polymorphism or sex on FA. However, a COMT effect on FA has been found to be moderated by sex in several regions. In particular, there were

| Volume-of-interest TBSS analysis
Using the VOI analysis to verify the results produced by the wholebrain TBSS analysis, we found out that only CC data survived this test. Kruskal-Wallis H-tests showed that, in males only, there was a statistically significant difference in FA values between the Val homozygotes and Met-carriers in the following regions of CC: genu and rostral midbody, anterior midbody, posterior midbody, and isthmus.  Figure 2).

| DISCUSSION
Our study demonstrated sexual dimorphism of FA values in various regions of the CC that was associated with COMT Val158Met polymorphism. In particular, FA values were higher in Val homozygous males than in Met-carriers males, whereas there was no genetic effect on FA values in females.
To the best of our knowledge, there have been only few studies directly investigating the relationships between COMT Val158Met polymorphism and white matter integrity.
Although Thomason et al. (2010) Kohannim et al. (2012) did not replicate these findings. The advantageous effect of Val homozygosity on white matter integrity has been reported by Liu et al. (2014), although in different brain regions.
Although we have not detected a main effect of genotype, we found that the COMT Val158Met polymorphism effect on FA in genu and rostral midbody of CC was moderated by sex, which adds an important dimension to the COMT functionality.
We have performed additional analysis to estimate potential effects of all variables of no-interest on the hypothesized interaction-to account for all covariates (i.e., sex × age, comt × age, age × sex) as proposed by Keller (2014).
The additional analysis supported our original results, that is, FA in male Val homozygotes was higher than in male Met-carriers in corpus callosum, SLF, IFOF, corona radiata, and CST at p < .05.
What are the implications of our finding of genotype-and sexrelated effects on CC fractional anisotropy? Corpus callosum contains WM fibers that connect the two hemispheres, and is involved in interhemispheric communication (Gazzaniga, 2000). Its known functions include interhemispheric exchange of sensory, motor, and cognitive information, and integration of inputs reaching one or both hemispheres (de Lacoste, Kirkpatrick, & Ross, 1985;Koch et al., 2011;Seltzer & Pandya, 1983;Wahl et al., 2007).
The literature of sex effect on CC size has been inconsistent.
A large sample study (Ardekani, Figarsky, & Sidtis, 2013) found that corpus callosum cross-sectional area was larger in females after correcting for the confounding effect of brain size (which is usually bigger in males). Another large cohort study (Prendergast et al., 2015), which used the same segmentation of CC as our group (Witelson, 1989), demonstrated main effects for sex when predicting CC length, perimeter, and circularity (i.e., measure of whole CC shape). In particular, males' callosa were longer than females, with greater perimeter while controlling for intracranial volume. The authors reported that greater absolute corpus callosum genu values were observed in males until the fifth decade compared with females. Importantly, in a DTI study, Westerhausen et al. (2011) have found the sex effect on genu of CC with FA in males being higher than in females.
Our finding of a sex-specific effect of COMT is in accordance with the literature on the role of COMT in sexual dimorphism. It has been known that estrogen reduces COMT enzyme activity, impacting on the effect of COMT genotype in women relative to men (Harrison & Tunbridge, 2008). Interestingly, sex-dichotomous effects of functional COMT genetic variations on cognitive functions disappear after menopause .
There is dearth of data on sex-specific effects of COMT genotype on brain white matter. An important study of Zinkstok et al. (2006) (Jacobs & D'Esposito, 2011). In light of our results of a sex-specific effect of COMT on brain white matter (CC), it would be interesting to further examine whether these discrepancies potentially might be explained by sex effect on COMT functionality.

| Limitations
Although our sample size was reasonably large, there were small numbers in some genotype groups, for example, Val homozygotes were represented by 12 males and 7 females. We have ensured that the Type 1 error has been controlled for by most rigorous two-stage statistics, that is, the whole-brain TBSS with subsequent nonparametric VOI analysis.
Another limitation of this study is that the individual plasma levels of estrogen were not assessed; therefore, we were unable to exclude any individual variation in estrogen levels. We assumed that in our group of healthy women (aged from 19 to 54), the levels of estrogen were within healthy range. Our conclusions on the inhibitory effect of estrogen on COMT activity have been based on existing literature.

| CONCLUSIONS
To the best of our knowledge, this is the first study to examine the effects of sex and COMT polymorphisms on white matter connectivity in healthy adults. We found sex-specific effect of COMT on