Genetic Vulnerability to Affective Psychopathology in Childhood: a Combined VBM and fMRI study

GCaAS

BHpOs

Mtd

R5f

Cdn

Kt

SvbasoGagtmCt

O

F

A

R

0d

enetic Vulnerability to Affective Psychopathology inhildhood: A Combined Voxel-Based Morphometrynd Functional Magnetic Resonance Imaging Study

ndrea Mechelli, Stefania Tognin, Philip K. McGuire, Diana Prata, Giuseppe Sartori, Paolo Fusar-Poli,tephane De Brito, Ahmad R. Hariri, and Essi Viding

ackground: The majority of affective psychopathology is rooted early in life and first emerges during childhood and adolescence.owever, little is known about how genetic vulnerability affects brain structure and function in childhood since the vast majority of studiesublished so far have been conducted on adult participants. The present investigation examined for the first time the effects of catechol--methyltransferase (COMT) valine (val) 158 methionine (met) (val158met) polymorphism, which has been shown to moderate predispo-

ition to negative mood and affective disorders, on brain structure and function in children.

ethods: Voxel-based morphometry and functional magnetic resonance imaging were used to measure gray matter volume and emo-ional reactivity in 50 children aged between 10 and 12 years. We tested the hypothesis that met158 allele affects structural brainevelopment and confers heightened reactivity within the affective frontolimbic circuit in children.

esults: The met158 allele was positively associated with gray matter volume in the left hippocampal head where genotype accounted for9% of interindividual variance. In addition, the met158 allele was positively associated with neuronal responses to fearful relative to neutral

acial expressions in the right parahippocampal gyrus where genotype accounted for 14% of the interindividual variance.

onclusions: These results indicate that the met158 allele is associated with increased gray matter volume and heightened reactivityuring emotional processing within the limbic system in children as young as 10 to 12 years of age. These findings are consistent with the

ey Words: Affective psychopathology, childhood, COMT, emo-ional processing, hippocampus

usceptibility for affective psychopathology depends on thedynamic interplay between genetic and environmental riskfactors (1,2). The catechol-O-methyltransferase (COMT)

aline (val) 158 methionine (met) (val158met) polymorphism haseen shown to moderate predisposition to negative mood andffective disorders. In recent years, several imaging genetictudies have demonstrated the effects of this and other risk genesn brain structure and function in healthy adult participants (3).iven that very few psychiatric illnesses arise de novo indulthood (4), it is important to extend the current imagingenetic work to include child samples. In the present investiga-ion, we used voxel-based morphometry (VBM) and functionalagnetic resonance imaging (fMRI) to examine the effects ofOMT val158met polymorphism on brain structure and emo-

ional processing in children aged between 10 and 12 years.Catechol-O-methyltransferase is an enzyme that catalyzes the

-methylation of extracellular dopamine in the brain (5) and is

rom the Department of Psychology (AM), Division of Psychological Medi-cine and Psychiatry (AM, ST, PKM, DP, PF-P), Social, Genetic and Devel-opmental Psychiatry Centre (DP, EV), and Forensic Mental Health Sci-ence Department (SDB), Institute of Psychiatry, King’s College London,London, United Kingdom; Department of Psychology (GS), University ofPadua, Padova, Italy; Department of Psychiatry (ARH), University of Pitts-burgh, Pittsburgh, Pennsylvania; and Division of Psychology and Lan-guage Sciences (EV), University College London, London, United King-dom.

ddress reprint requests to Andrea Mechelli, Ph.D., PO Box 67, Institute ofPsychiatry, King’s College London, 103 Denmark Hill, London, SE5 8AF,United Kingdom; E-mail: [email protected].

eceived September 3, 2008; revised January 30, 2009; accepted January 30,

006-3223/09/$36.00oi:10.1016/j.biopsych.2009.01.033

mainly found in its membrane-bound (MB-COMT) form inpostsynaptic neurons (6,7). The most abundant expression ofCOMT, both in terms of messenger RNA density (7,8) andenzyme activity (8,9), is found in the prefrontal cortex and theparahippocampal gyrus. The enzymatic activity of COMT ismodulated by a guanine (G) to adenine (A) single nucleotidepolymorphism (SNP) change (known as val158met or rs4680) inthe COMT gene. This translates into a valine to methionineamino acid change in codon 158 that causes a threefold tofourfold decrease in its molecular thermostability. The alleleshave been shown to be codominant with the met158 alleleassociated with decreased COMT activity, resulting in highersynaptic dopamine levels; the val158 allele associated withincreased COMT activity, resulting in lower synaptic levels; andthe heterozygote genotype (val158/met158) associated with anintermediate level of COMT activity (8).

Several studies suggest that the met158 allele is advantageousfor cognitive performance (10–15) and prefrontal function(11,15–18) not only in adults but also in children (19,20).However, a series of recent studies have also implicated themet158 allele in negative mood and affective disorders, includingincreased levels of anxiety in women, obsessive-compulsivedisorder in men, panic disorder, alcoholism, aggressiveness,bipolar affective disorder, major depression, and higher sensitiv-ity to pain (as reviewed by Drabant et al. [21]). The met158 allelehas also been associated with a high level of anxiety (22) andearly-onset antisocial behavior (23) in children and adolescents,although a recent investigation of emotional symptoms in chil-dren 6 to 7 years old did not find an association (24).

While the impact of the val158met polymorphism on prefron-tal function has been characterized extensively in recent years,only a few functional imaging studies have explored the rela-tionship between this polymorphism and brain activation during

BIOL PSYCHIATRY 2009;66:231–237© 2009 Society of Biological Psychiatry.

Published by Elsevier Inc. All rights reserved.

dan(vmapmbtitdrs

vuiee(abcf

pemasmmiyfpm

tCcpatow

M

S

pToctpA((

232 BIOL PSYCHIATRY 2009;66:231–237 A. Mechelli et al.

w

ose-dependent increase in limbic and prefrontal activationssociated with met158 during the processing of unpleasant butot pleasant visual stimuli in a small group of healthy volunteersn � 35). More recently, Drabant et al. (21), using 101 healthyolunteers exposed to fearful and angry facial expressions, foundet158 to be associated with a dose-dependent increase in

ctivation within a frontolimbic circuit including the hippocam-us and the ventrolateral prefrontal cortex; furthermore, inet158 homozygotes, there was increased functional couplingetween limbic and prefrontal regions. These studies suggesthat met158 increases reactivity within a frontolimbic circuit thats critical for emotional regulation, thereby providing support tohe implication of this allele in negative mood and affectiveisorders. The same frontolimbic circuit appears to show COMT-elated differences in gray matter volume as revealed by recenttructural neuroimaging studies (28–30).

It is currently unclear, however, whether the effects of theal158met polymorphism on brain responses to emotional stim-li that have been reported in adult participants are also evidentn children. This is an important question since there is increasingvidence that the majority of affective psychopathology is rootedarly in life and first emerges during childhood and adolescence4). If gene-related differences in adult participants reflect alter-tions that occurred during childhood and adolescence, then aetter characterization of these alterations during childhood isritical for understanding how genes affect brain structure andunction to mediate vulnerability to affective psychopathology (31).

It is also unclear whether variation in the val158met polymor-hism is associated with differences in brain morphology fromarly age. Dopaminergic innervation increases during brainaturation and decreases during late adolescence and early

dulthood (32). Thus, it has been proposed that differences inynaptic dopamine levels associated with the val158met poly-orphism may be associated with different trajectories of brainaturation (30). A recent investigation demonstrated that the

mpact of COMT genotype on gray and white matter density inoung adults is dependent on the age of the participants inemale subjects but not in male subjects (30). However, norevious studies have examined whether the val158met poly-orphism affects structural brain development in childhood.We therefore used VBM and fMRI to examine for the first time

he impact of the functional val158met polymorphism in theOMT gene on brain structure and emotional processing in 50hildren aged between 10 and 12 years. Participants wereresented with pictures of fearful and neutral faces on a screennd were required to detect the gender of each face. We testedhe hypothesis that met158 allele affects structural brain devel-pment and confers increased sensitivity to emotional stimuliithin the affective frontolimbic circuit in children.

ethods and Materials

ubjectsA total of 50 boys aged 10 to 12 years old participated in the

resent study. Participants were recruited from the longitudinalwins Early Development Study (TEDS) database as part of anngoing twin neuroimaging project that included mostly typi-ally developing children, as well as an oversample of children inhe top 10% of the United Kingdom population for conductroblems (Supplement 1). The short version of the Wechslerbbreviated Scales of Intelligence (WASI) was used to assess IQ33). In addition, the Strengths and Difficulties Questionnaire

ww.sobp.org/journal

problems, hyperactivity, peer problems, prosociality, and totalbehavioral difficulties in all participants as rated by both parentsand teachers. All subjects were genotyped for the val158met inthe COMT gene (see below). Our sample of 50 volunteerscomprised 14 met158/met158, 22 val158/met158, and 14 val158/val158 individuals. A one-way analysis of variance (ANOVA)revealed that the three genotype groups did not differ in age, IQ,or any of the SDQ indicators (p � .05) (Table 1).

Experimental TaskThe experimental paradigm involved presenting emotional

and neutral faces taken from the Pictures of Facial Affect (35) butcropped to remove hair. For each face, subjects had to make agender classification (male or female) by pressing left or rightresponse buttons; no explicit recognition or categorization of theemotional expression was required. Each stimulus was presentedfor 3000 msec and successive stimuli were separated by aninterstimulus interval of 750 msec, resulting in a stimulus onsetasynchrony of 3750 msec. A total of 80 stimuli were presented ona computer screen in a single scanning session that lasted 6minutes and 24 seconds; the stimuli were arranged in 10 blocks,each comprising eight fearful or eight neutral faces. The exper-imental paradigm also comprised two “rest” blocks in which nofaces were presented but a fixation cross remained on the screenfor 32 seconds. The order of presentation of fearful and neutralfaces was counterbalanced across subjects.

GenotypingDNA was extracted from blood or cheek swabs using stan-

dard methods (36). Genotyping of the rs4680, which encodesthe val158met polymorphism, was performed by KBioscience(http://www.kbioscience.co.uk; Hertz, United Kingdom) using acompetitive allele-specific polymerase chain reaction (PCR) sys-tem (CASP). The region amplified was atcacccagcggatggtggatt-tcgctggc[A/G]tgaaggacaaggtcacccttgtggttggag. The genotypingresults of a sample of 130 subjects, which included our 50participants, were under Hardy-Weinberg equilibrium.

Image AcquisitionStructural brain images were acquired using a General Electric

Signa 3.0 Telsa Excite II magnetic resonance imaging (MRI)scanner (General Electric Medical Systems, Milwaukee, Wiscon-sin) at the Institute of Psychiatry. Structural scanning cons-isted of an isotropic resolution three-dimensional (3-D) inversionrecovery prepared spoiled gradient echo. Two hundred through-plane partitions (each 1.1 mm thick) were collected, with twopartitions being discarded at each end of the imaging volume tominimize wrap-round artefacts.

In addition, functional image volumes (192 scans for eachsubject) were collected using T2*-weighted gradient echo-planarimaging (EPI) sequence with 28 slices (slice thickness 3.5 mm,gap � .3 mm) covering the whole brain (repetition time [TR] �2 sec, echo time [TE] � 25 msec, field of view � 220 � 220,matrix size 64 � 64). Stimuli were projected onto a high-resolution screen located in front of the participant’s head andwere viewed via a mirror attached to the head coil.

Data AnalysisBehavioral Data. Analysis of response accuracy and reaction

times was performed using the Statistical Package for SocialScience (SPSS), version 15.0 (SPSS Inc., Chicago, Illinois). Thethree genotype groups were compared using a 2 � 3 ANOVAwith facial expression as repeated measures. Inferences were

cwubtsIfaaawwva�(aaCs

wuKesc1fr8(

T

nAFT

P

%%AA

d(

quotQ

A. Mechelli et al. BIOL PSYCHIATRY 2009;66:231–237 233

Voxel-Based Morphometry. Structural images were prepro-essed using optimized voxel-based morphometry implementedith Statistical Parametric Mapping software (SPM5) runningnder Matlab 7.0 (Mathworks, Sherbon, Massachusetts). Voxel-ased morphometry is a whole-brain, unbiased, semi-automatedechnique for characterizing regional cerebral differences intructural magnetic resonance images (Supplement 2) (37–39).nferences were made using a statistical threshold of p � .05 afteralse discovery rate (FDR) correction for multiple comparisonscross the whole brain and an extent threshold of 10 voxels. Inddition to the whole-brain analysis, we performed an additionalnalysis with a mask that comprised six regions of interest (ROIs)ith a radius of 6 mm (total size of the mask: 266). These ROIsere selected on the basis that they showed a significant effect ofal158met on brain structure in a recent investigation (28) withdult participants and included the left (x � �17, y � �1, z �25) and right (x � 19, y � �14, z � �17) hippocampus, the left

x � �28, y � 1, z � �23) and right (x � 29, y � 2, z � �23)mygdala, the left inferior frontal gyrus (x � �32, y � 32, z � 5),nd the right superior frontal gyrus (x � 15, y � 33, z � 32).oordinates are reported in Montreal Neurological Institute (MNI)pace.

Statistical Parametric Mapping. Functional image volumesere analyzed using SPM5 software (http//www.fil.ion.cl.ac.uk/spm; Functional Imaging Laboratory, London, Unitedingdom), running under Matlab 7.0 (Supplement 3). Significantffects at group level (40) were identified using t contrasts and atatistical threshold of p � .05 after FDR correction for multipleomparisons across the whole brain and an extent threshold of0 voxels. . In addition to the whole-brain analysis, we per-ormed an additional analysis with a mask that comprised twoegions of interest with a radius of 6 mm (total size of the mask:8). These ROIs were selected because in a recent investigation

able 1. Participants’ Characteristics and Task Performance

Met/Met Met/V

14 22ge in Months 137.8 � 9.2 135.2 �ull Scale IQ 104.3 � 11.9 102.0 �eacher Rated SDQ Scores

Hyperactivity 3.8 � 2.6 3.7 �Conduct problems 1.6 � 1.5 1.9 �Emotional problems 1.9 � 2.0 1.0 �Prosociality 8.4 � 1.7 8.3 �Peer problems .8 � 1.4 1.5 �Total difficulties 8.1 � 5.2 7.7 �

arent Rated SDQ ScoresHyperactivity 3.9 � 4.2 4.2 �Conduct problems 1.3 � 2.3 1.9 �Emotional problems 1.3 � 1.3 1.2 �Prosociality 6.5 � 2.8 6.3 �Peer problems .8 � 1.2 1.6 �Total difficulties 7.2 � 5.7 8.9 �Correct Responses Neutral 96.8 � 3.5 93.7 �Correct Responses Fear 96.3 � 3.3 93.7 �

verage RT Neutral (msec) 953.84 � 233.49 1082.17 �verage RT Fear (msec) 937.90 � 225.53 998.97 �

Data are expressed as mean values (� standard deviation). p values foifferent genotypes; p values for % correct responses and average RT refe

Neutral, Fear) as repeated measure.Met, methionine; Val, valine; ANOVA, analysis of variance; IQ, intelligence

uestionnaire.

21) they showed heightened reactivity to emotional facial

expressions in adult participants with the met/met variant com-pared with those with the val/val variant and included the rightparahippocampal gyrus (x � 15, y � �33, z � �7) and the rightventrolateral prefrontal cortex (x � 55, y � 22, z � 13).Coordinates are reported in MNI space.

Functional Connectivity Analysis. Functional connectivity isa measure of the temporal correlation of the blood oxygenationlevel-dependent (BOLD) signal in spatially remote regions and isused widely in the imaging community as a simple and robustcharacterization of aspects of functional integration. A previousinvestigation (21) found that, during the processing of emotionalfacial expressions, the functional coupling between limbic andprefrontal regions was increased in adult participants with themet/met variant compared with those with the val/val variant.We therefore examined whether a similar effect of COMTval158met on functional integration could be detected within ourcohort of children using psychophysiological interaction (PPI)(Supplement 4) (41). Inferences were made at group level (40)using a statistical threshold of p � .05 after FDR correction formultiple comparisons across the whole brain and an extentthreshold of 10 voxels; since no significant effects were detectedusing this statistical threshold, we report trends significant at p �.05 (uncorrected) for completeness.

Results

Behavioral PerformanceTable 1 reports the response accuracy and reaction times of

each genotypic group for neutral and fearful facial expressionsindependently. Catechol-O-methyltransferase genotype was notsignificantly associated with percentage of correct responses(p � .063) or reaction times (p � .319); in addition, there was no

Val/Val All Subjects p Value

14 50140.6 � 9.6 137.4 � 9.5 .251106.5 � 9.1 103.9 � 11.7 .534

1.9 � 2.5 3.2 � 2.6 .0941.4 � 1.2 1.7 � 1.5 .6231.1 � 1.1 1.3 � 1.5 .2308.4 � 1.4 8.3 � 1.6 .9521.3 � 2.7 1.2 � 2.4 .7355.6 � 5.0 7.2 � 5.5 .442

2.4 � 2.7 3.6 � 3.2 .2331.3 � 2.1 1.5 � 2.3 .7161.0 � 1.9 1.1 � 1.8 .9347.0 � 2.6 6.6 � 2.8 .7871.0 � 2.4 1.2 � 2.2 .5725.7 � 6.1 7.4 � 6.9 .411

96.0 � 4.1 95.1 � 6.6 .06396.4 � 4.1 95.3 � 6.0

03 1000.67 � 213.65 1023.43 � 210.16 .31930 1022.41 � 163.21 981.63 � 167.80

, full scale IQ, and SDQ scores refer to a one-way ANOVA contrasting the2 � 3 ANOVA contrasting the different genotypes with facial expression

ient; RT, reaction time; n, number of subjects; SDQ, Strengths and Difficulties

al

9.813.0

2.51.61.31.82.86.0

2.62.62.03.22.68.17.97.9

184.141.

r ager to a

significant interaction between COMT genotype and facial ex-

www.sobp.org/journal

pr

E

pbirecm�rcazEgh

ejpap�s

E

swipcgse

E

ab

F(wtcv

234 BIOL PSYCHIATRY 2009;66:231–237 A. Mechelli et al.

w

ression on percentage of correct responses (p � .834) oreaction times (p � .362).

ffect of COMT Genotype on Gray Matter VolumeThere were no significant differences in gray matter volume at

� .05 (corrected for multiple comparisons across the wholerain). However, when using small volume correction to exam-ne six regions of interest where an effect of COMT had beeneported in adult participants (24), we detected two significantffects (Figure 1). Individuals with the met158/met158 genotypeompared with val158/val158 individuals expressed greater grayatter volume in the left hippocampal head (x � �24, y � 4, z �24; Z score � 2.7; p � .026 corrected; cluster size: 33). A linear

egression analysis in statistical parametric mapping (SPM) indi-ated that gray matter volume in this region was positivelyssociated with the number of met158 alleles (x � �24, y � 4,� �24; Z score � 2.6; p � .036 corrected; cluster size: 26).stimate of the R2 measure in SPSS revealed that val158metenotype accounted for 59% of interindividual variance in the leftippocampal head.

We considered the possibility that the COMT-related differ-nces in gray matter volume might be associated with intersub-ect differences in IQ, level of emotional problems, conductroblems, or hyperactivity. When these variables were includeds covariates of no interest, the effect of COMT val158metolymorphism in the left hippocampal head was replicated (x �24, y � 4, z � �24; Z score � 2.6; p � .042 corrected; cluster

ize: 15).

ffect of Facial Expression on Brain ActivationThe processing of fearful stimuli compared with neutral

timuli (i.e., fearful � neutral facial expressions) was associatedith increased activation in a distributed bilateral network

ncluding the prefrontal cortex, the insula, the amygdala, thearahippocampal gyrus, the thalamus, the superior parietalortex, the fusiform gyrus, and the middle and anterior cingularyrus (Figure 2). In contrast, there were no regions showingignificant activation for neutral compared with fearful facialxpressions.

ffect of COMT Genotype on Brain ActivationThere were no areas showing an effect of genotype significant

t p � .05 (corrected for multiple comparisons across the whole

igure 1. Effects of COMT val158met polymorphism on gray matter volumep � .05 after FDR correction). Sagittal view of the left hippocampal head

here gray matter volume was greater for met/met relative to val/val geno-ype; parameter estimates are shown for each genotypic group. COMT,atechol-O-methyltransferase; FDR, false discovery rate; met, methionine;al, valine.

rain). However, when using small volume correction to exam-

ww.sobp.org/journal

ine our two regions of interest based on the results of a similarstudy with adult participants (see Methods and Materials), wedetected a significant effect in the right parahippocampal gyrus(Figure 3). In this region, individuals with the met158/met158genotype showed greater activation than those with the val158/val158 genotype in response to fearful relative to neutral facialexpressions (x � 14, y � �32, z � 0; Z score � 2.8; p � .028corrected; cluster size: 114). The met158/val158 individualsexpressed intermediate values between the two homozygotegroups; a linear regression analysis in SPM confirmed that brainactivation in the right paraparahippocampal gyrus was positivelyassociated with the number of met158 alleles (x � 14, y � �32,z � 6; Z score � 3.1; p � .019 corrected; cluster size: 117).Estimate of the R2 measure in SPSS revealed that COMT genotypeaccounted for 14% of interindividual variance in BOLD signal. Incontrast, we did not replicate the previous finding of greateractivation in met158/met158 relative to val158/val158 individualsin the ventrolateral prefrontal cortex of adult participants.

We examined whether intersubject differences in IQ, level ofemotional problems, conduct problems, or hyperactivity mightaccount for the COMT-related difference in the right hippocam-pal formation. When these variables were included as covariatesof no interest, the effect in the right hippocampal formation wasreplicated (x � 14, y � �32, z � 0; Z score � 2.5; p � .045corrected; cluster size: 53).

Functional ConnectivityWe used psychophysiological interaction to investigate the

changes in functional integration that mediated heightened acti-

Figure 2. Regions activated for fearful relative to neutral facial expressionson average across the three genotype groups (p � .05 after FDR correction).

vsww2scpivtfa

D

tsIsec

pvpwaf

FFamgyoalm

A. Mechelli et al. BIOL PSYCHIATRY 2009;66:231–237 235

ation in the right parahippocampal gyrus. No effects reachedignificance after correction for multiple comparisons; however,hen lowering the statistical threshold to p � .05 (uncorrected),e detected a trend in the medial anterior cingulate cortex (x �, y � 34, z � �6; Z score � 2.8; p � .003 uncorrected; clusterize: 233) (Figure 3). This area showed increased functionaloupling with the right parahippocampal gyrus during therocessing of fearful relative to neutral facial expressions in

ndividuals with the met/met homozygotes compared with val/al homozygotes. Estimate of the R2 measure in SPSS revealedhat genotype accounted for 15% of interindividual variance inunctional integration between the right parahippocampal gyrusnd the medial anterior cingulate cortex.

iscussion

The aim of the present study was to investigate for the firstime the impact of the COMT val158met polymorphism on braintructure and function in children aged between 10 and 12 years.n particular, we tested the hypothesis that met158 allele affectstructural brain development and confers increased sensitivity tomotional stimuli within the affective frontolimbic circuit inhildren.

Using optimized voxel-based morphometry, we replicated theositive association between the met158 allele and gray matterolume in the left hippocampal head that had been found in arevious investigation with adult participants (28). In contrast,e did not replicate the positive association between the val158llele and gray matter volume in prefrontal regions (28). These

igure 3. Effects of COMT val158met on emotional processing (p � .05 afterDR correction). Top: the right parahippocampal gyrus expressed increasedctivation in response to fearful relative to neutral facial expression in met/et relative to val/val individuals; parameter estimates are shown for each

enotypic group. Bottom: an analysis of psychophysiological interactionielded a trend (p � .003 uncorrected) indicating that increased activationf the right parahippocampal gyrus in met/met relative to val/val individu-ls was mediated by increased functional coupling with the anterior cingu-

ate cortex. COMT, catechol-O-methyltransferase; FDR, false discovery rate;et, methionine; val, valine.

indings indicate that at least some of the COMT-related differ-

ences in brain structure that have been described in the adultpopulation are already evident by the age of 10 to 12 years. Inaddition, they suggest a possible dissociation between the limbicsystem, where the effect of our polymorphism of interest wasfound in our cohort of children, and prefrontal regions, where noeffects were detected even when lowering the statistical thresh-old to p � .05 (uncorrected). The lack of COMT-related differ-ences in the prefrontal cortex can be considered surprising, giventhat the COMT enzyme alters extracellular dopamine levels in theprefrontal cortex and that its activity increases from neonate intoadulthood (42,43).

Using functional magnetic resonance imaging, we detectedincreased activation in the right parahippocampal gyrus duringemotional processing in met/met relative to val/val homozy-gotes. The parahippocampal gyrus shows abundant expressionof COMT, both in terms of messenger RNA density (7,8) andenzyme activity (8,9); this provides support to the idea thatmet158-associated increases in dopamine level might underlieheightened reactivity of this region (21). Neuroimaging studieswith human participants and animal lesion models provideconverging evidence that the parahippocampal gyrus is part of afrontolimbic circuit that mediates anxious states and behaviorsand is critical for emotional regulation (44–47). This has led tothe suggestion that heightened reactivity to emotional stimuli inthe right parahippocampal gyrus of individuals with the met158allele might be associated with increased sensitivity to negativeenvironmental cues (21). The absence of COMT-related differ-ences in task performance confirms that the alteration in the rightparahippocampal gyrus is likely to reflect differences in theimplicit processing of the emotional content of the stimuli ratherthan the gender discrimination task itself. Our finding replicatesthe results of a previous study using the same experimentalparadigm with adult participants (21). However, the presentinvestigation expands these results by demonstrating for the firsttime that the COMT val158met polymorphism affects emotionalprocessing in children. This is an important observation, sincereports of an association between the met158 allele and negativemood/affective disorders have typically been based on adultparticipants; yet, the majority of affective psychopathology doesnot arise de novo in adults without any warning in childhood oradolescence (4). Our finding that the met158 allele is associatedwith increased sensitivity to negative environmental cues in theparahippocampal gyrus of children as young as 10 to 12 years ofage is consistent with the notion that genetic factors affect brainfunction to moderate vulnerability to affective psychopathologyfrom early age (31).

It may be considered surprising that we found no evidence foran effect of COMT genotype on amygdala activation, given theimplication of this region in emotional processing (48). Never-theless, human postmortem studies indicate that COMT is mini-mally expressed in the amygdala (49). A previous imaginggenetic study reported an association between the met158 alleleand amygdala reactivity to unpleasant and pleasant pictures (27);however, amygdala activation did not vary as a function ofCOMT genotype in another imaging genetic study that usedemotional and nonemotional faces (21). These inconsistentresults might be explained by differences in task design (i.e.,biologically salient arousal vs. discriminating valence of complexvisual scenes).

To examine the neuronal interactions that mediated theimpact of the COMT val158met polymorphism in the rightparahippocampal gyrus, we performed an analysis of psycho-

www.sobp.org/journal

cttradtsmc(iorc

glpceTcprtfmc

iarbtfitIarbt(mp

owatpSiwpefpvftr

236 BIOL PSYCHIATRY 2009;66:231–237 A. Mechelli et al.

w

ance after correction for multiple comparisons, we found arend in the medial anterior cingulate cortex. This indicatedhat heightened reactivity during emotional processing in theight parahippocampal gyrus of met158 homozygotes wasssociated with heightened functional coupling with the me-ial anterior cingulate cortex. Clearly, this finding must beaken with caution and requires replication in future studiesince it did not reach significance at p � .05 after correction forultiple comparisons. Nevertheless, we note that the anterior

ingulate cortex has been implicated in emotional regulation50), as well as negative mood and affective disorders (51). Its also noteworthy that the genotype effect accounted for 15%f interindividual variance in functional coupling between theight parahippocampal gyrus and the medial anterior cingulateortex.

In the present investigation, the effects of COMT val158metenotype on gray matter volume and emotional processing wereocalized in the left hippocampal head and the right parahip-ocampal gyrus, respectively. Only one previous study hadombined structural and functional techniques to examine theffect of COMT genotype on brain structure and function (52).hat study, which used a working memory paradigm and wasonducted with adult participants at high genetic risk for schizo-hrenia, revealed that the val158 allele was associated with botheduced gray matter density and increased regional activation ofhe anterior cingulate cortex. In contrast, we found no evidenceor an association between COMT-related differences in grayatter volume and COMT-related differences in emotional pro-

essing (Supplement 5).We also report that val158met heterozygotes expressed an

ntermediate gray matter volume in the left hippocampal headnd showed intermediate reactivity in the analyses of regionalesponses and functional connectivity. A similar load effect haseen found in previous neuroimaging studies that investigatedhe impact of the COMT val158met polymorphism on prefrontalunction (11–15) and emotional reactivity (21,26,27). This patterns consistent with the results of in vitro thermostability studieshat have shown that the two alleles act codominantly (53).nterestingly, the effects of COMT genotype on brain structurend function were localized in the left and right hemisphere,espectively. The left-lateralization of the structural effect mighte due to limited statistical power, given that a previous inves-igation with a larger adult sample reported bilateral differences28). In contrast, the right-lateralization of the functional effectight reflect the use of faces as stimuli, since faces are known toreferentially activate the right hemisphere (54).

The present investigation has a number of limitations. First,ur cohort only included male subjects and therefore it is unclearhether the results can be generalized to the female population;n increasing number of animal and human studies indicate thathe effect of the COMT val158met polymorphism on brainhysiology may be gender-dependent (see [55] for review).econd, the sample size was relatively small for a neuroimagingnvestigation of gene-related effects but was nevertheless in lineith that of several imaging genetic studies. Our sample of 50articipants was sufficient for detecting statistically significantffects of our polymorphism of interest on brain structure andunction, which replicated previous findings with adult partici-ants. Third, a more comprehensive understanding of howariation in the COMT val158met polymorphism affects brainunction in response to emotional stimuli will require investiga-ion of interactions with other candidate genes and with envi-

ww.sobp.org/journal

In conclusion, this is the first investigation to examine theeffect of the COMT val158met polymorphism on brain structureand function within a sample of children. We found that themet158 allele is associated with increased gray matter volume ofthe left hippocampal head and heightened reactivity of theparahippocampal gyrus during emotional processing in childrenas young as 10 to 12 years of age. These results suggest that themet158 allele affects structural brain development and confersincreased sensitivity to emotional stimuli within the limbic sys-tem before the possible manifestation of any symptoms inadolescence and adulthood. To establish whether the genotypiceffects reported in the present investigation are related to clinicalsymptomatology later in life, a longitudinal approach will berequired.

This work was supported by grants from the Medical ResearchCouncil (United Kingdom) (G0401170 to EV) and the Depart-ment of Health Forensic Mental Health Programme (MRD 12-73to EV).

Professor Robert Plomin’s support for this research has beenvital. We also thank Dr. Alice Jones, Mr. Andrew McMillan, andMrs. Patricia Busfield for their role in the recruitment and datacollection.

The authors report no biomedical financial interests or po-tential conflicts of interest.

Supplementary material cited in this article is availableonline.

1. Moldin SO, Gottesman II (1997): At issue: Genes, experience, and chancein schizophrenia–positioning for the 21st century. Schizophr Bull 23:547–561.

2. Caspi A, Moffitt TE (2006): Gene-environment interactions in psychiatry:Joining forces with neuroscience. Nat Rev Neurosci 7:583–590.

3. Fisher PM, Muñoz KE, Hariri AR (2008): Identification of neurogeneticpathways of risk for psychopathology. Am J Med Genet C Semin MedGenet 148:147–153.

4. Kim-Cohen J, Caspi A, Moffitt TE, Harrington H, Milne BJ, Poulton R(2003): Prior juvenile diagnoses in adults with mental disorder: Devel-opmental follow-back of a prospective-longitudinal cohort. Arch GenPsychiatry 60:709 –717.

5. Saito S, Iida A, Sekine A, Miura Y, Sakamoto T, Ogawa C, et al. (2001):Identification of 197 genetic variations in six human methyltranferasegenes in the Japanese population. J Hum Genet 46:529 –537.

6. Napolitano A, Cesura AM, Da PM (1995): The role of monoamine oxidaseand catechol-O-methyltransferase in dopaminergic neurotransmission.J Neural Transm Suppl 45:35– 45.

7. Matsumoto M, Weickert CS, Akil M, Lipska BK, Hyde TM, Herman MM, etal. (2003): Catechol-O-methyltransferase mRNA expression in humanand rat brain: Evidence for a role in cortical neuronal function. Neuro-science 116:127–137.

8. Chen J, Lipska BK, Halim N, Ma QD, Matsumoto M, Melhem S, et al.(2004): Functional analysis of genetic variation in catechol-O-methyl-transferase (COMT): Effects on mRNA, protein, and enzyme activity inpostmortem human brain. Am J Hum Genet 75:807– 821.

9. Masuda M, Tsunoda M, Imai K (2003): High-performance liquid chroma-tography-fluorescent assay of catechol-O-methyltransferase activity inrat brain. Anal Bioanal Chem 376:1069 –1073.

10. Bruder GE, Keilp JG, Xu H, Shikhman M, Schori E, Gorman JM, et al. (2005):Catechol-O-methyltransferase (COMT) genotypes and working mem-ory: Associations with differing cognitive operations. Biol Psychiatry 58:901–907.

11. Egan MF, Goldberg TE, Kolachana BS, Callicott JH, Mazzanti CM, StraubRE, et al. (2001): Effect of COMT Val108/158 Met genotype on frontallobe function and risk for schizophrenia. Proc Natl Acad Sci U S A 98:6917– 6922.

12. Malhotra AK, Kestler LJ, Mazzanti C, Bates JA, Goldberg T, Goldman D(2002): A functional polymorphism in the COMT gene and performance

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

3

3

A. Mechelli et al. BIOL PSYCHIATRY 2009;66:231–237 237

3. Goldberg TE, Egan MF, Gscheidle T, Coppola R, Weickert T, KolachanaBS, et al. (2003): Executive subprocesses in working memory: Relation-ship to catechol-O-methyltransferase Val158Met genotype and schizo-phrenia. Arch Gen Psychiatry 60:889 – 896.

4. Diamond A, Briand L, Fossella J, Gehlbach L (2004): Genetic and neuro-chemical modulation of prefrontal cognitive functions in children. Am JPsychiatry 161:125–132.

5. Caldú X, Vendrell P, Bartrés-Faz D, Clemente I, Bargalló N, Jurado MA, etal. (2007): Impact of the COMT Val108/158 Met and DAT genotypes onprefrontal function in healthy subjects. Neuroimage 37:1437–1444.

6. Bertolino A, Caforio G, Blasi G, De Candia M, Latorre V, Petruzzella V, et al.(2004): Interaction of COMT (Val(108/158)Met) genotype and olanzap-ine treatment on prefrontal cortical function in patients with schizo-phrenia. Am J Psychiatry 161:1798 –1805.

7. Bertolino A, Caforio G, Petruzzella V, Latorre V, Rubino V, Dimalta S, et al.(2006): Prefrontal dysfunction in schizophrenia controlling for COMTVal158Met genotype and working memory performance. Psychiatry Res147:221–226.

8. Blasi G, Mattay VS, Bertolino A, Elvevåg B, Callicott JH, Das S, et al. (2005):Effect of catechol-O-methyltransferase val158met genotype on atten-tional control. J Neurosci 25:5038 –5045.

9. Diamond A, Briand L, Fossella J, Gehlbach L (2004): Genetic and neuro-chemical modulation of prefrontal cognitive functions in children. Am JPsychiatry 161:125–132.

0. Wahlstrom D, White T, Hooper CJ, Vrshek-Schallhorn S, Oetting WS,Brott MJ, Luciana M (2007): Variations in the catechol-O-methyltrans-ferase polymorphism and prefrontally guided behaviors in adolescents.Biol Psychiatry 61:626 – 632.

1. Drabant EM, Hariri AR, Meyer-Lindenberg A, Munoz KE, Mattay VS,Kolachana BS, et al. (2006): Catechol-O-methyltransferase val158metgenotype and neural mechanisms related to affective arousal and reg-ulation. Arch Gen Psychiatry 63:1396 –1406.

2. Olsson CA, Anney RJ, Lotfi-Miri M, Byrnes GB, Williamson R, Patton GC(2005): Association between the COMT Val158Met polymorphism andpropensity to anxiety in an Australian population-based longitudinalstudy of adolescent health. Psychiatr Genet 15:109 –115.

3. Thapar A, Langley K, Fowler T, Rice F, Turic D, Whittinger N, et al. (2005):Catechol-O-methyltransferase gene variant and birth weight predictearly-onset antisocial behavior in children with attention-deficit/hyper-activity disorder. Arch Gen Psychiatry 62:1275–1278.

4. Evans J, Xu K, Heron J, Enoch MA, Araya R, Lewis G, et al. (2008): Emo-tional symptoms in children: The effect of maternal depression, lifeevents, and COMT genotype. Am J Med Genet B Neuropsychiatr Genet150B:209 –218.

5. Bishop SJ, Cohen JD, Fossella J, Casey BJ, Farah MJ (2006): COMT geno-type influences prefrontal response to emotional distraction. Cogn Af-fect Behav Neurosci 6:62–70.

6. Smolka MN, Schumann G, Wrase J, Grüsser SM, Flor H, Mann K, et al.(2005): Catechol-O-methyltransferase val158met genotype affects pro-cessing of emotional stimuli in the amygdala and prefrontal cortex.J Neurosci 25:836 – 842.

7. Smolka MN, Bühler M, Schumann G, Klein S, Hu XZ, Moayer M, et al.(2007): Gene-gene effects on central processing of aversive stimuli. MolPsychiatry 12:307–317.

8. Cerasa A, Gioia MC, Labate A, Liguori M, Lanza P, Quattrone A (2008):Impact of catechol-O-methyltransferase Val(108/158) Met genotype onhippocampal and prefrontal gray matter volume. Neuroreport 19:405–408.

9. Taylor WD, Züchner S, Payne ME, Messer DF, Doty TJ, MacFall JR, et al.(2007): The COMT Val158Met polymorphism and temporal lobe mor-phometry in healthy adults. Psychiatry Res 155:173–177.

0. Zinkstok J, Schmitz N, van Amelsvoort T, de Win M, van den Brink W,Baas F, et al. (2006): The COMT val158met polymorphism and brainmorphometry in healthy young adults. Neurosci Lett 405:34 –39.

1. Viding E, Williamson DE, Hariri AR (2006): Developmental imaging ge-netics: Challenges and promises for translational research. Dev Psycho-pathol 18:877– 892.

2. Rosenberg DR, Lewis DA (1995): Postnatal maturation of the dopami-nergic innervation of monkey prefrontal and motor cortices: A tyrosine

33. Wechsler D (1999): Wechsler Abbreviated Scale of Intelligence. San Antonio,TX: Psychological Corporation.

34. Goodman R, Ford T, Simmons H, Gatward R, Meltzer H (2000): Using theStrengths and Difficulties Questionnaire (SDQ) to screen for child psy-chiatric disorders in a community sample. Br J Psychiatry 177:534 –539.

35. Ekman P, Friesen WV (1976): Pictures of Facial Affect. Palo Alto, CA: Con-sulting Psychologists Press.

36. Freeman B, Smith N, Curtis C, Huckett L, Mill J, Craig IW (2003): DNA frombuccal swabs recruited by mail: Evaluation of storage effects on long-term stability and suitability for multiplex polymerase chain reactiongenotyping. Behav Genet 33:67–72.

37. Ashburner J, Friston KJ (2000): Voxel-based morphometry–the meth-ods. Neuroimage 11:805– 821.

38. Good CD, Johnsrude IS, Ashburner J, Henson RN, Friston KJ, FrackowiakRS (2001): A voxel-based morphometric study of ageing in 465 normaladult human brains. Neuroimage 14:21–36.

39. Mechelli A, Price CJ, Friston KJ, Ashburner J (2005): Voxel-based mor-phometry of the human brain: Methods and applications. Curr MedImaging Rev 1:105–113.

40. Penny WD, Holmes AP (2003): Random effects analysis. In: FrackowiakRSJ, Friston KJ, Frith C, Dolan R, Friston KJ, Price CJ, et al., editors. HumanBrain Function. San Diego, CA: Academic Press, 843– 850.

41. Friston KJ, Buechel C, Fink GR, Morris J, Rolls E, Dolan RJ (1997): Psycho-physiological and modulatory interactions in neuroimaging. Neuroim-age 6:218 –229.

42. Levitt P (2003): Structural and functional maturation of the developingprimate brain. J Pediatr 143(suppl 4):S35–S45.

43. Tunbridge EM, Weickert CS, Kleinman JE, Herman MM, Chen J,Kolachana BS, et al. (2007): Catechol-o-methyltransferase enzyme activ-ity and protein expression in human prefrontal cortex across the post-natal lifespan. Cereb Cortex 17:1206 –1212.

44. Rusch BD, Abercrombie HC, Oakes TR, Schaefer SM, Davidson RJ (2001):Hippocampal morphometry in depressed patients and control subjects:Relations to anxiety symptoms. Biol Psychiatry 50:960 –964.

45. Ploghaus A, Narain C, Beckmann CF, Clare S, Bantick S, Wise R, et al.(2001): Exacerbation of pain by anxiety is associated with activity in ahippocampal network. J Neurosci 21:9896 –9903.

46. McNaughton N, Gray JA (2000): Anxiolytic action on the behaviouralinhibition system implies multiple types of arousal contribute to anxi-ety. J Affect Disord 61:161–176.

47. McNaughton N (1997): Cognitive dysfunction resulting from hippocam-pal hyperactivity–a possible cause of anxiety disorder? Pharmacol Bio-chem Behav 56:603– 611.

48. Kim H, Somerville LH, Johnstone T, Polis S, Alexander AL, Shin LM,Whalen PJ (2004): Contextual modulation of amygdala responsivity tosurprised faces. J Cogn Neurosci 16:1730 –1745.

49. Hong J, Shu-Leong H, Tao X, Lap-Ping Y (1998): Distribution of catechol-O-methyltransferase expression in human central nervous system. Neu-roreport 9:2861–2864.

50. Posner MI, Rothbart MK, Sheese BE, Tang Y (2007): The anterior cingu-late gyrus and the mechanism of self-regulation. Cogn Affect BehavNeurosci 7:391–395.

51. Pavuluri MN, O’Connor MM, Harral EM, Sweeney JA (2008): An fMRIstudy of the interface between affective and cognitive neural circuitry inpediatric bipolar disorder. Psychiatry Res 162:244 –255.

52. McIntosh AM, Baig BJ, Hall J, Job D, Whalley HC, Lymer GK, et al. (2007):Relationship of catechol-O-methyltransferase variants to brain struc-ture and function in a population at high risk of psychosis. Biol Psychiatry61:1127–1134.

53. Weinshilboum RM, Otterness DM, Szumlanski CL (1999): Methylationpharmacogenetics: Catechol-O-methyltransferase, thiopurine methyl-transferase, and histamine N-methyltransferase. Annu Rev PharmacolToxicol 39:19 –52.

54. Ishai A, Schmidt CF, Boesiger P (2005): Face perception is mediated by adistributed cortical network. Brain Res Bull 67:87–93.

55. Harrison PJ, Tunbridge EM (2007): Catechol-O-methyltransferase(COMT): A gene contributing to sex differences in brain function, and tosexual dimorphism in the predisposition to psychiatric disorders. Neu-

www.sobp.org/journal