ORIGINAL RESEARCH ARTICLE

Catechol-O-methyl transferase Val158Met genepolymorphism in schizophrenia: working memory, frontallobe MRI morphology and frontal cerebral blood flowB-C Ho1, TH Wassink1, DS O’Leary1, VC Sheffield2 and NC Andreasen1,3,4

1Department of Psychiatry, Roy J and Lucille A Carver College of Medicine, University of Iowa, Iowa City, IA, USA;2Department of Pediatrics, Division of Molecular Genetics, Howard Hughes Medical Institute, University of Iowa, Iowa City, IA,USA; 3The MIND Institute, Albuquerque, NM, USA; 4Department of Psychiatry, University of New Mexico, Albuquerque,NM, USA

The catechol-O-methyl transferase (COMT) gene is considered a leading schizophreniacandidate gene. Although its role in increasing schizophrenia susceptibility has beenconflicting, recent studies suggest the valine allele may contribute to poor cognitive functionin schizophrenia. V158M COMT genotype was obtained on 159 schizophrenia patients and 84healthy controls. The effects of COMT genotype on four measures of working memory/executive functions (Wisconsin Card Sorting, digit span backward, Trail Making and N-backtests) and on MRI frontal brain volumes were examined. Genotype distributions were notsignificantly different between patients and controls. There were no significant genotype orgenotype-by-group effects on any working memory/executive function measures. No genotypeor genotype-by-diagnosis interaction effects were found with MRI frontal lobe volumes.Randomization analyses using [15O]H2O positron emission tomography (PET) cerebral bloodflow data found Val/Val patients had higher frontal lobe activation than Met/Met patients whileperforming the one-back task. Overall, these findings do not support a major role for COMT inincreasing susceptibility for schizophrenia or in mediating frontal lobe function. Age-relatedchanges and phenotypic heterogeneity of schizophrenia may influence the complex relation-ships between COMT genotype and cognition.Molecular Psychiatry (2005) 10, 287–298. doi:10.1038/sj.mp.4001616Published online 25 January 2005

Keywords: endophenotype; dopamine; MRI; candidate gene; association study; PET

Schizophrenia is a severe and disabling neuropsychiatricdisorder. Although a century of genetic epidemiologicalresearch has provided strong evidence for a genetic basis,success in the search for schizophrenia susceptibilitygenes has thus far been limited. Risks for developingschizophrenia appear to be related to complex interac-tions between multiple genes, and between genes andenvironmental factors. The phenotype of schizophreniais equally complex, and encompasses deficits in multipledomains of brain function, thereby presenting an elusiveand difficult target for research into the neurobiologicaland genetic underpinnings of schizophrenia.1,2

One strategy to address the challenges from geneticcomplexity and phenotypic heterogeneity has been tostudy endophenotypes.3,4 Endophenotypes can be neu-rocognitive, neurophysiological, biochemical or neuroa-natomical traits that correlate with the disease and/ordisease severity.5 These traits are heritable and can bemeasured in unaffected relatives. Studying endopheno-

types (vs using psychiatric diagnostic categories) maypotentially provide more power for genetic studies.6,7

Endophenotypes are presumably biologically closer tothe gene, transmitted in a less complex manner and canbe more accurately assessed. Using such an approach,investigations into putative endophenotypes, such assmooth-pursuit eye movements and prepulse inhibi-tion, have already identified potential schizophreniasusceptibility genetic loci and candidate genes.8–11

Working memory is widely considered to be anendophenotype for schizophrenia. Working memoryrefers to the neuropsychological construct of ‘tempor-ary storage and manipulation of the informationnecessary for such complex cognitive tasks as lan-guage comprehension, learning and reasoning’.12

Schizophrenia patients as well as their unaffectedrelatives perform poorly on tasks that assess workingmemory function.13–23 Working memory is moderatelyheritable with 43–49% of its variability accounted bygenetic factors24 In a series of studies involving apopulation isolate, a significant additive geneticheritability estimate was found in verbal workingmemory, and for visual working memory at a trendlevel.20,21,25 Spatial working memory performance wasmore highly correlated in monozygotic twins than in

Received 18 May 2004; revised 9 August 2004; accepted 4 October2004

Correspondence: Dr B-C Ho, Department of Psychiatry 2880 JPP,University of Iowa, 200 Hawkins Drive, Iowa City, IA 52252, USA.E-mail: beng-ho@uiowa.edu

Molecular Psychiatry (2005) 10, 287–298& 2005 Nature Publishing Group All rights reserved 1359-4184/05 $30.00

www.nature.com/mp

dizygotic twin pairs. Greater familial loading wasassociated with greater impairment in working mem-ory performance. Working memory also appears tobe a stable trait. Deficits have been demons-trated throughout the course of schizophrenia,26,27 inneuroleptic-naı̈ve, medicated and unmedicatedpatients, and are relatively unaffected by antipsy-chotic treatment.16,28–30

Dopamine neurotransmission in the dorsolateralprefrontal cortex (DLPFC) plays a pivotal role inworking memory. Single unit recording studies innon-human primates as well as functional neuroima-ging studies in humans have consistently demon-strated this relationship.31,32 The mesocorticaldopamine system, which ascends from the ventraltegmental area, provides extensive connections to theneocortex,33,34 where D1 receptors are the most highlyexpressed of the dopamine receptors.35,36 In non-human primates studied using delayed responseparadigms, stimulation of D1 receptors in the DLPFCenhances working memory.37,38 These observations,together with the association between workingmemory deficits and aberrant DLPFC activation inschizophrenia patients, provide convergent evidencethat disruptions in DLPFC dopamine neurotran-smission may underlie working memory deficits inschizophrenia.29,39–43

Elimination of extracellular dopamine relies onsynaptic reuptake (via dopamine transporter andnorepinephrine transporter) as well degradation bymonoamine oxidase and catechol-O-methyl transfer-ase (COMT).44–47 Owing to its role in the catabolism ofdopamine, the COMT gene has been extensivelystudied as a schizophrenia candidate gene. In hu-mans, there is a common V158M functional poly-morphism48 where a methionine (Met) substituted fora valine (Val) leads to a four-fold reduction in COMTenzyme activity.49 Furthermore, the COMT gene islocalized to chromosome 22q11, a region which hasbeen implicated in schizophrenia linkage studies.50

Deletion of chromosome 22q11 (as seen in velocar-diofacial syndrome) is associated with high rates ofschizophrenia-like psychosis.51 A recent study hasgenerated renewed interest in this gene, especiallywith regard to its role in mediating working memoryand cognition.52 The association between low-activityMet allele and better cognitive performance may berelated to increased prefrontal dopamine bioavail-ability.52,53 However, even though the COMT gene is

an appealing candidate gene in schizophrenia, case–control as well as family-based association studies todate have had mixed findings with regard to itscontribution to schizophrenia susceptibility.52,54–63

The aim of this study is to examine the role ofV158M COMT polymorphism in working memory inschizophrenia. This study integrates data from multi-modal investigational approaches, including neurop-sychology, molecular genetics, morphometric andfunctional neuroimaging, to probe the genetic andphenotypic complexities of schizophrenia. This studyextends the growing literature on COMT by examin-ing several concurrent measures of working memory/executive function in a sample of recent-onsetschizophrenia patients (as compared to the morechronically ill patients examined in previous stu-dies). In addition, no prior studies have examined therelationships between V158M COMT genotype andMRI frontal lobe morphology, nor have they comparedprefrontal cerebral blood flow between schizophreniapatients with different COMT genotypes.

Materials and methods

SubjectsIn all, 159 schizophrenia patients and 84 healthyvolunteers were obtained through the University ofIowa Mental Health Clinical Research Center(MHCRC). These subjects have participated in variousMHCRC research studies approved by the Universityof Iowa institution review board. All subjects gavewritten informed consent to undergo research assess-ments. Patients were evaluated using a semistruc-tured interview instrument, ComprehensiveAssessment of Symptoms and History (CASH),64 fromwhich schizophrenia diagnoses meeting DSM-III-R orDSM-IV criteria were based. Healthy volunteers wererecruited from the community through newspaperadvertisements. They were initially screened bytelephone, and further evaluated using an abbreviatedversion of the CASH to exclude subjects with currentor past medical, neurological or psychiatric illnesses.

To minimize the likelihood of spurious associationarising from ethnic stratification, only subjects ofCaucasian ancestry were included into this report.Sociodemographic characteristics of the sample aresummarized in Table 1. Both groups were of compar-able age. A significantly greater proportion of patientswere male. Patients also had fewer years of education

Table 1 Sociodemographics of the sample

Controls Patients Statistic P-value

N (% males) 84 (40.48) 159 (74.21) w2¼ 26.7 o0.001Age (years) 27.0 (7.03) 26.5 (7.35) t241¼ 0.53 0.60Education (years) 14.8 (1.55) 12.9 (1.94) t241¼ 8.51 o0.0001Full scale IQ 109 (12.05) 91.0 (12.45) t241¼ 11.01 o0.0001Parental education (years) 13.6 (2.18) 13.7 (2.63) t241¼ 0.49 0.63

COMT and frontal lobe functions and morphology in schizophreniaB-C Ho et al

288

Molecular Psychiatry

and lower full scale IQ. However, parental educa-tional attainment was comparable between thegroups. Patients were evaluated relatively early inthe course of their illness. The mean duration ofillness (since the onset of first psychiatric hospitali-zation) was 2.5 years (SD¼4.73). A total of 79patients (48%) were evaluated during their firstpsychiatric hospitalization, and 57 (35%) were neu-roleptic naı̈ve. The median duration of lifetimeneuroleptic treatment was 2 months (25th and 75thinterquartile range¼ 12 months).

Genetic analysesDNA was prepared by high-salt extraction from wholeblood.65 The COMT codon 108/158 polymorphismwas determined by restriction fragment length poly-morphism analysis. PCR amplification was performedon a 165 bp fragment of the COMT gene using theprimers 50-GGGCCTACTGTGGCTACTCA-30 (forward)and 50-GGCCCTTTTTCCAGGTCTGACA-30 (back-ward). Each 20 ml PCR reaction contained 20 nggenomic DNA, 5.0 pM of each primer, 400mM of eachdNTP, 0.1 U of Taq DNA polymerase, 2.0ml PCR buffer(100 mM Tris-HCl (pH 8.8), 500 mM KCl, 15 mMMgCl2, 0.01% gelatin (w/v)), 4.0 ml betaine and10.5 ml water. Cycling conditions were as follows:initial denaturation was at 941C for 3 min, followed by40 temperature cycles consisting of 30 s at 941C fordenaturation, 30 s at 601C for annealing and 30 s at721C for extension, followed by a final elongation at721C for 5 min. The PCR products were then digestedby restriction enzyme NlaIII at 371C for 3 h. Digestionproducts were electrophoresed on 4% NuSieveagarose gel, and visualized by ethidium bromidestaining under ultraviolet light. Genotyping wasperformed blind to subject’s clinical status.

Working memory/executive function assessmentAlthough working memory performance has beenwidely studied, there remains considerable debatewith regard to its definition and how to best quantifythe construct.66 Some investigations emphasize thetemporary storage component of working memory,while others stress the importance of measuring theexecutive function component of working memory (ieboth holding information ‘online’ as well as manip-ulation of stored information). Until a better con-sensus has been reached with regard to its definitionand measurement, working memory performancemay be best assessed by using a combination ofdifferent tests.66,67

In this study, working memory assessment consistsof a battery of four tests: Wisconsin Card Sorting Test(WCST),68 WAIS-R digit span backward, Trail Makingtest and N-back test. The first three working memorytests are part of a standard neuropsychological testbattery, while the N-back test was administeredduring a positron emission tomography (PET) studyon a subset of the sample (see below). WCST raw testscores for the number of perseverative errors werenormalized for age and educational attainment, and

transformed to t-scores based on population norms68

(population mean¼ 50, SD¼10, higher scores indi-cate better performance). The WAIS-R digit spansubtest typically combines forward and backwardperformance into a single score. In this study, only thetotal raw score of the number of correct items from thebackward portion of the test is used, since previousstudies suggest that the backward span task is a morerobust measure of working memory.17,66 The TrailMaking test is a measure of executive functioning andset alternation.69 Trail A involves connecting numbersin succession, while Trail B requires subjects toconnect numbers and letters alternately in successiveorder. Working memory load in the Trail Making testconsists of keeping the next target in mind whilesearching for the target on the sheet. By subtractingTrail A from Trail B, it is possible to assess frontal lobefunctions by eliminating motor speed.70 The timedifference (s) taken to connect all items correctly onTrails A and B is used in the analysis.

MRI acquisition and image processingImages were obtained on a 1.5-T GE Signa MRscanner. Three different MR sequences were acquiredfor each subject (ie T1-weighted spoiled grass, protondensity (PD) and T2-weighted images). The imagingparameters have been described previously.71 Theimages were processed using the locally developedBRAINS (Brain Research: Analysis of Images, Net-works, and Systems) software package. Detaileddescriptions of image analysis methods have beenprovided elsewhere.71–74 In brief, the T1-weightedimages were spatially normalized and resampled sothat the anterior–posterior axis of the brain wasrealigned parallel to the anterior–posterior commis-sure line, and the interhemispheric fissure wasaligned on the other two axes. The T2- and PD-weighted images were aligned to the spatially normal-ized T1-weighted image using an automated imageregistration program.75 These images were then sub-jected to a linear transformation into standardizedstereotaxic Talairach atlas space76 to generate auto-mated measurements of frontal, temporal, parietal,and occipital lobes, cerebellum, and subcorticalregions.77 To further classify tissue volumes into graymatter (GM), white matter (WM) and cerebrospinalfluid (CSF), we employed a discriminant analysismethod of tissue segmentation based on automatedtraining class selection that utilized data from the T1,T2 and PD sequences.78 In this study, we examinedfrontal lobe GM, WM and CSF volumes.

PET data acquisition and image processingDuring image acquisition, subjects lay reclined andare oriented in the PET scanner with laser light guidesaligned at the orbital meatal line of the brain. Stimulifor the one-back task were presented on a videomonitor. The baseline condition consisted of acheckerboard flashing at a rate of 0.5 Hz for 40 s.Subjects were told to fixate on a dot at the center ofthe checkerboard. During the one-back condition, the

COMT and frontal lobe functions and morphology in schizophreniaB-C Ho et al

289

Molecular Psychiatry

same 0.5 Hz flashing checkerboard stimulus waspresented but each checkerboard had an asterisk.The location of the asterisk changed with each flash.The subject was instructed to press the index fingerbutton if he/she saw the asterisk appear in the samelocation on two consecutive flashes. Although sub-jects were told that this may happen more than once,only one set of two consecutive flashes were pre-sented 35 s into the stimulus. Reaction time (ms) inthis one-back task was used for analysis.

Regional cerebral blood flow (rCBF) was measuredusing the bolus [15O]H2O method79 with a GE4096-PLUS Scanner. A total of 15 slices (6.5 mm center-to-center) were acquired with a 10-cm axial field of view.Images were acquired over a 100-s interval followingvenous injection of 50–75 mCi of [15O]H2O. Imageswere reconstructed for a 40-s interval following bolustransit, determined by time–activity curves from aregion-of-interest over a cerebral artery.80 Arterialblood sampling allowed calculation of tissue perfu-sion in ml/min/100 g tissue using the autoradio-graphic method.80 The outlines of the PET imageswere automatically identified with an edge detectionalgorithm and the PET images for each condition foreach subject were coregistered with their MR imagesusing a variance minimization program.75 An 18 mmHanning filter was applied to the PET images for eachcondition to eliminate residual anatomical variability.

The PET study was carried in 17 healthy volunteersand 16 schizophrenia patients to examine rCBFdifferences during working memory performance.Among the 17 healthy controls, two subjects hadMet/Met genotype, 12 Met/Val and three Val/Val. Ofthe 16 schizophrenia patients, seven had Met/Metgenotype, three Met/Val and six Val/Val. Owing tothis genotype distribution, only the seven Met/Metand six Val/Val patients were used in subsequent PETanalyses.

Statistical analyses

COMT genotype frequencies were compared betweensubjects and controls using w2 tests. Analyses of therelationships between COMT genotype and the threemeasures of working memory/executive function (ieWCST, WAIS-R digit span backward and Trail Makingtest) and with MRI frontal lobe volumes (GM, WMand CSF) were performed using ANOVAs. In eachgeneral linear model, the respective trait measure wasentered as the dependent measure with genotype andaffectation status as independent variables. Intracra-nial volume and age were entered as covariates in theMRI frontal lobe morphology analyses. For the[15O]H2O PET data, a nonparametric randomizationanalysis81 was performed to compare the rCBFresponse of seven Met/Met patients and six Val/Valpatients on the one-back task. This randomizationanalysis relies on across-task and across-group com-parisons for one-back condition minus baselinecondition in Met/Met patients as opposed to one-back condition minus baseline condition in Val/Val

patients. We used an uncorrected alpha of 0.005 inthe PET randomization analysis.

Results

Genotype frequency distribution of the sample, andworking memory/executive functioning and MRIfrontal lobe volumes broken down by COMT geno-type, are summarized in Table 2. Met allele frequencywas 0.52 for the whole sample. Genotype distribu-tions did not deviate from Hardy–Weinberg expecta-tions (w2¼ 0.50 and 0.95 in healthy volunteer andpatient groups respectively, df¼ 2, P’sZ0.62).Although a greater proportion of schizophreniapatients had the Val/Val genotype (19.5 vs 14.2% incontrols), the genotype frequency distributions werenot significantly different between the two samples(w2¼2.56, P¼ 0.28). There were no significant geno-type or genotype-by-group effects on age, educationalattainment or full scale IQ (F’sr1.27, df¼2,243,P’sZ0.27).

Effects of COMT genotype on working memory/executive functionsThe effects of COMT genotype on WCST, digit spanbackwards and Trail Making tests were examinedusing ANCOVA analyses. Patients had significantlylower WCST perseverative errors t-scores thanhealthy volunteers (mean¼ 52.3 (SD¼13.56) vs 56.6(SD¼ 11.10); main effect of group F¼ 5.14, df¼1,243,P¼0.02). There were no significant genotype orgenotype-by-group effects on WCST performance(Table 2 and Figure 1; F’s¼1.74 and 0.32, respec-tively, df¼ 2,243, P’sZ0.18). Similarly, patients per-formed more poorly on digit span (mean’s¼ 5.98(SD¼ 1.77) vs 8.12 (SD¼ 2.20) in healthy volunteers)and on Trail Making test (mean’s¼ 49.2 (SD¼36.06)and 28.7 (SD¼15.03), respectively; main effect ofgroup F’sZ20.8, df¼ 1,200 or 1,211, P’so0.0001).There were also no significant genotype or genotype-by-group effects on digit span or Trail Makingperformance (Table 2 and Figure 1; F’sr0.53,df¼ 2,200 or 2,211, P’sZ0.59).

Since the patient group had a significantly greaterpreponderance of male subjects, fewer years ofeducation and lower full scale IQ than healthyvolunteers, these three variables were entered asadditional covariates in the general linear models.The results remained unchanged, with no signi-ficant genotype or genotype-by-group effects on anyof the three measures of working memory/executivefunctioning.

Effects of COMT genotype on MRI frontal lobemorphologyIn all, 49 of the 84 healthy volunteers and 100 of the168 patients had MRI frontal lobe volumetric data.The remaining 35 healthy volunteers and 68 patientsdid not have available MRI scans for a variety ofreasons, including inability to undergo scanningprocedure because of claustrophobia, poor quality

COMT and frontal lobe functions and morphology in schizophreniaB-C Ho et al

290

Molecular Psychiatry

scan

sor

scan

sobta

ined

usin

gn

on

com

para

ble

ima-

gin

gp

ara

mete

rs(e

ither

on

an

old

er

or

new

er

scan

nin

gp

roto

col).

Th

ere

were

no

statistic

ally

sign

ifican

td

iffere

nces

betw

een

health

yvolu

nte

ers

with

an

dw

ithou

tM

Rd

ata

with

regard

togen

der

com

positio

n,

CO

MT

gen

oty

pe

distrib

utio

nor

age

(deta

ilsof

statistic

al

an

aly

ses

availa

ble

up

on

requ

est).

Patie

nts

with

MR

data

were

also

not

sign

ifican

tlyd

iffere

nt

from

those

with

ou

tM

Rd

ata

on

avarie

tyof

socio

-d

em

ogra

ph

ican

dclin

ical

measu

res:

gen

der

com

posi-

tion

,C

OM

Tgen

oty

pe

distrib

utio

n,

age

an

dd

ura

tion

of

illness

or

sym

pto

mse

verity

(deta

ilsof

statistic

al

an

aly

ses

availa

ble

up

on

requ

est).

Th

em

ean

fron

tal

lobe

GM

,W

Man

dC

SF

volu

mes

(ad

juste

dfo

rin

tracra

nia

lvolu

me

an

dage)

for

49

health

yvolu

nte

ers

an

d100

sch

izop

hre

nia

patie

nts

bro

ken

dow

nby

CO

MT

gen

oty

pe

are

sum

mariz

ed

inT

able

2an

dF

igu

re2.

Th

ere

were

no

statistic

ally

sign

ifican

teffe

cts

of

gen

oty

pe

on

fron

tal

lobe

mor-

Table 2 Mean (SD) working memory/executive functioning and frontal lobe MRI volumes (cm3) in controls and schizophrenia patients, and ANOVA analyses (F statistic(P)) on the main effects of diagnostic grouping, genotype and genotype-by-grouping (total brain compartment volume and age as covariates)

Controls Patients Diagnosis F(P)

Genotype F(P)

Genotype byDiagnosis F (P)

Met/Met Met/Val Val/Val Met/Met Met/Val Val/Val

N 18 54 12 38 90 31

Working memory/executive functionsa

WCST 53.3 (10.75) 56.7 (11.01) 60.8 (11.42) 50.7 (14.36) 52.5 (13.94) 53.7 (11.47) 5.14 (0.02) 1.74 (0.18) 0.32 (0.73)Digit span (backwards) 8.3 (2.03) 8.2 (2.79) 7.6 (2.15) 5.9 (1.88) 5.9 (2.03) 6.2 (2.08) 30.2 (o0.0001) 0.12 (0.89) 0.53 (0.59)Trail Making test 28.4 (11.80) 28.9 (14.68) 28.2 (21.18) 54.7 (42.70) 45.0 (29.32) 52.0 (44.32) 20.8 (o0.0001) 0.34 (0.71) 0.46 (0.63)

Frontal lobe volumes (cm3)GM 277.9 (14.24) 268.1 (13.67) 257.4 (15.54) 279.6 (16.51) 266.5 (17.33) 270.6 (11.49) 1.98 (0.16) 1.06 (0.35) 2.08 (0.13)WM 184.9 (14.16) 177.5 (16.24) 174.1 (15.85) 184.3 (15.60) 174.4 (16.60) 173.8 (13.95) 0.20 (0.65) 0.58 (0.56) 0.08 (0.92)CSF 32.2 (14.32) 34.5 (15.53) 45.0 (28.07) 36.5 (16.31) 45.4 (18.06) 38.9 (23.43) 0.37 (0.55) 1.43 (0.24) 2.58 (0.08)

aWCST: Wisconsin Card Sorting Test (t-scores).WAIS-R digit span, backwards (number of correct, total raw score).Trail Making test (Trail B minus Trail A (s)).

Figu

re1

Work

ing

mem

ory

/execu

tive

fun

ctio

ns

perfo

r-m

an

ce

inh

ealth

yvolu

nte

ers

an

din

sch

izop

hre

nia

patie

nts

bro

ken

dow

nby

CO

MT

gen

oty

pe.

COM

Tand

frontallobefunctions

andm

orphologyin

schizophreniaB

-CH

oet

al

291

Mo

lecu

lar

Psy

chia

try

phology (F’so1.43, df¼ 2,148, P’s40.24). No signifi-cant genotype by group interaction effects on frontalGM or WM (F’so2.08, df¼ 2,148, P’s40.13). Geno-type-by-group interaction effects approached but didnot achieve statistical significance for frontal CSFvolume (F’s¼2.58, df¼ 2,148, P¼ 0.08).

Effects of COMT genotype on [15O]H2O rCBF duringN-back task performanceAll seven schizophrenia patients with Met/Metalleles and six patients with Val/Val alleles were ableto perform the one-back version of the N-back task.Each subject responded correctly by pressing thebutton when the asterisk appeared at the same

location on two consecutive flashes of the checker-board. Met/Met allele patients had slightly longerreaction times than Val/Val patients, but this differ-ence was not statistically significant (mean’s¼ 1382.4and 1248.5 ms, respectively; t¼0.18, df¼ 10.2,P¼0.86). Randomization analyses indicated thatthere were no regions within the DLPFC or withinthe anterior cingulate where Met/Met patientsshowed significantly greater rCBF than Val/Valpatients. Table 3 and Figure 3 summarize brainregions where schizophrenia patients with Met/Metalleles had significantly lower rCBF than patientswith Val/Val alleles. While performing the one-backtask, Val/Val patients showed greater activation in theleft dorsal lateral prefrontal cortex (Brodmann’s area47) and in the right mesial frontal lobe (Brodmann’sareas 8 and 9).

Discussion

In this study, we found that V158M COMT genotypefrequencies did not differ significantly betweenschizophrenia patients and healthy volunteers. Wealso investigated the effects of COMT genotype on awide range of frontal lobe-related measures. Workingmemory/executive function performance (in WCST,digit span backwards and Trail Making tests) and MRIfrontal lobe volumes (GM, WM and CSF) were notsignificantly different between genotype groupings.Patients with Val/Val genotype showed greater activa-tion in the frontal lobes than Met/Met patients whileperforming the N-back task. Although this differencein rCBF, based on a small subset of the sample, may betaken to mean that subjects with valine allele haveprefrontal cortical inefficiency, the overall findingsfrom this study do not support a major role for COMTin increasing susceptibility for schizophrenia, or inmediating frontal lobe structure and functions.

V158M COMT functional polymorphism is probablythe most widely investigated candidate gene inschizophrenia research. The bulk of the evidence todate suggests that this polymorphism confers, at best,a small increase in susceptibility to the disorder. Mostcase–control studies, as well as our study, have failedto find a significant association between COMTpolymorphism and schizophrenia.52,54,55,59–61,82–88 Ahandful of case–control studies found that the highenzyme activity valine allele had a modest effect onconferring increased risk for schizophrenia.58,63,89

Conversely, the methionine allele has also beenreported to occur more frequently among schizophre-nia patients than normal controls.90,91 Family-basedstudies have had equally mixed results, with somestudies finding preferential transmission of the valineallele in schizophrenia patients52,56,57 while othershave detected no significant associations.92–94

These conflicting results with regard to the role ofV158M COMT polymorphism in schizophrenia sus-ceptibility may relate to inadequate statistical powerof individual association studies or to variability inassociation between different ethnic populations.

Figure 2 Mean frontal lobe GM, WM and CSF volumes forhealthy volunteers and for schizophrenia patients brokendown by COMT genotype.

COMT and frontal lobe functions and morphology in schizophreniaB-C Ho et al

292

Molecular Psychiatry

When data from case–control association studies werepooled, the two published meta-analyses indicate thatV158M COMT allele variation does not contribute toincreased vulnerability for schizophrenia.62,95 Family-based studies, on the other hand, suggest that thevaline allele may be a risk factor in schizophreniapatients of European ancestry.62 To clarify the truerole of the V158M COMT polymorphism in schizo-

phrenia susceptibility, future studies will requirelarger sample sizes and will need to minimize theeffects of population stratification.

Beyond its potential as a susceptibility gene, therehas also been a long-standing interest within schizo-phrenia research with regard to the phenotypiccorrelates of the COMT gene. The methionine allelehas been associated with aggression, suicide, worseprognosis and greater positive symptom severity inschizophrenia.91,96–100 Egan et al’s finding of anassociation between the valine allele and poor WCSTperformance has generated renewed interest in thisarea of investigation, especially with regard to theeffects of COMT in mediating cognition in schizo-phrenia.52 This finding has been replicated in part byJoober et al, who found that genotype effects onWCST performance among schizophrenia patientswas at a trend level of significance, and wasnonsignificant in healthy volunteers.59 In anotherstudy, healthy volunteers with the valine allelecommitted significantly more WCST perseverativeerrors than those with methionine allele.101 Theeffects of V158M COMT polymorphism appear toextend beyond WCST performance, and have beenimplicated in other measures of prefrontal neurocog-nitive functions (such as N-back task102 and prefrontalP300 amplitude and latency103,104), and in mediatingother cognitive domains (such as processing speedand attention105 and ‘cognitive stability and flexibil-ity’106). All these studies have found an associationbetween the valine allele and poorer cognitiveperformance, and postulate that the underlyingmechanism may be related to lower prefrontaldopamine levels arising from higher dopamine cata-bolism mediated by the valine allele.

Table 3 Regions where schizophrenia patients with Met/Met alleles had significantly lower rCBF than patients with Val/Valalleles while performing the one-back task

Brain regiona Talairach coordinates Significance of tmax (P)b Number of voxelsc

x y z

Right cerebellum 36 �52 �36 �3.58 (0.0005) 1.2Left DLPFC (Brodmann 47) �39 20 �3 �3.20 (0.0014) 0.4Right cerebellum 11 �61 �36 �3.11 (0.0018) 0.3Left precuneus (Brodmann 30) �10 �55 19 �3.28 (0.0011) 0.3Right mesial frontal lobe (Brodmann 8) 5 23 31 �3.03 (0.0022) 0.2Right mesial frontal lobe (Brodmann 8) 9 53 13 �3.11 (0.0018) 0.2Left posterior–inferior lobe cerebellum �32 �67 �40 �3.21 (0.0014) 0.2Right middle temporal gyrus 47 �30 6 �3.18 (0.0015) 0.1

aRegions were named from the inspection of coregistered magnetic resonance and PET images, as well as Talairachcoordinates. These coordinates represent spatial orientation with respect to a point located in a horizontal plane through theanterior and posterior commissures (z¼ 0), at the midline of the brain slice (x¼ 0) and at the anterior commissure line (y¼ 0).The x-axis was the distance in millimeters to the left (negative) and to the right (positive) of the midline. The y-axis was thedistance in millimeters anterior (positive) or posterior (negative) to the anterior commissure line. The z-axis was the distancein millimeters above (positive) or below (negative) a horizontal plane through the anterior and posterior commissures.btmax is the highest t-value identified in the peak.cNumber of voxels in the peaks that exceeded alpha¼ 0.005.

Figure 3 Regions where cerebral blood flow among Met/Met schizophrenia patients were significantly lower thanthose in Val/Val patients while performing the one-backtask.

COMT and frontal lobe functions and morphology in schizophreniaB-C Ho et al

293

Molecular Psychiatry

In contrast to these previous studies, we did notfind a significant association between the highenzyme activity valine allele and poor WCST perfor-mance. In fact, among schizophrenia patients as wellas healthy volunteers, none of the working memory/executive function measures we examined differedsignificantly across genotype groupings. Our findingsof no significant genotype effects are consistent with arecent study involving 200 healthy volunteers.107

Fossella et al found that V158M COMT genotypestatus did not predict performance in any of the fourmeasures of attention they examined. Interestingly,healthy volunteers with alleles expected to result inhigher synaptic dopamine actually performed morepoorly in the attention tasks. The four-repeat allele ofthe DRD4 exon III polymorphism and 10-repeat alleleof the DAT 30-VNTR polymorphism were bothsignificantly associated with poorer executive atten-tion scores. Although the COMT Met allele showedno significant associations, subjects with Met/Metgenotype had numerically worse attention scores.The MAOA-LPR three-repeat allele, which wouldalso be expected to result in higher synaptic dopa-mine levels, showed a trend towards poorer attentionperformance.

The younger age, milder cognitive impairment and/or shorter illness duration of our sample maypotentially explain our nonreplication of the associa-tion between the valine allele and poor cognitiveperformance. Except for the Tsai et al study104 (whichstudied Han Chinese female nursing students agedbetween 19 and 21 years), the mean age of subjects inprevious studies52,59,101–103,105,106,108 are substantiallyolder than ours. Age-related loss in frontal lobedopamine function has been repeatedly demonstratedin post-mortem as well as in vivo neuroimagingstudies (eg Rinne,109 de Keyser et al,110 Suhara etal,111 Kaasinen et al112 and Volkow et al113). Thisdecline in dopamine function appears to beginearly,114 and has been estimated to occur at a rate of11–13% per decade in age.112 Thus, even thoughsubjects from previous studies52,59,101–103,105,106,108 werestill in their mid-adulthood, and were only 10–15years older than our sample, these older subjects mayalready have considerably lower dopamine functionalreserve than our younger subjects. Any furtherreduction in dopamine function among these olderindividuals, such as an increased dopamine catabo-lism associated with the presence of valine allele,may have more pronounced detrimental effects onfrontal lobe functions than in younger individuals.

Compared with the Egan et al and Bilder et alstudies,52,105 which examined similar neuropsycholo-gical measures, our sample was cognitively lessimpaired and our patients less chronically ill. Themean WCST t-scores for patients and controls in Eganet al study were 37.6 (712.6) and 49.4 (79.0),respectively52 (vs 52.3 (712.2) and 56.6 (79.6) inour patients and controls subjects, respectively). Inthe Bilder et al study,105 the mean number ofperseverative errors among schizophrenia patients

was 49.2 (vs 11.2 in our patients). The mean illnessduration among patients in the Bilder et al study andin our study is 19.3 and 2.5 years, respectively.Although the Egan et al study did not report illnessduration, the average length of illness would likely be10–12 years based on the mean age of the subjects.

These differences in severity of cognitive impair-ment and in duration of illness may be indicative ofgreater heterogeneity in our patient sample. Eventhough all three studies utilized DSM diagnosticcriteria to ascertain patients, it is well recognizedthat there can be substantial variability in patientswho share the same set of clinical symptoms ofschizophrenia. Since our patients consisted of amixture of first episode, recent-onset and chronicallyill schizophrenia patients, our sample may be morerepresentative of the ‘average’ schizophrenia patient.However, this heterogeneity will likely add ‘noise’ tothe analyses, and may be less ideal for geneticassociation studies. On the other hand, the Egan etal and Bilder et al samples, being more chronically illand having greater cognitive impairment, may begenetically more homogeneous, and therefore, morelikely to detect effects of COMT genotype on cognitiveperformance.

A third reason for our lack of significant associa-tions between COMT genotype and working memoryperformance may be related to the secondary role ofCOMT in determining synaptic dopamine levels.There is limited human data with regard to therelative contributions of catecholamine (dopamineand norepinephrine) transporters, monoamine oxi-dase and COMT in influencing synaptic dopaminelevels in the prefrontal cortex. However, animalstudies indicate that COMT plays only a minimalrole in the clearance of synaptic dopamine, evenunder conditions of low dopamine transporter activ-ity.45–47 Furthermore, COMT genotype appears toaccount for only a small amount (2–4%) of sharedvariance with WCST performance.52,105 Thus, thissmall effect of COMT genotype on cognitive perfor-mance may be especially difficult to detect in youngerand more heterogeneous sample of schizophreniapatients.

To the best of our knowledge, this is the first studythat has examined MRI brain morphometric correlatesof COMT genotype. We did not find any significantassociations between genotype and frontal GM, WMor CSF volumes. Although our group has previouslyreported associations between BDNF and NOTCH4with these same measures of frontal lobe,115,116 futurestudies on the relationships between COMT and brainmorphology will need to examine specific regionswithin the frontal lobes, especially regions known tomediate working memory.

Our finding of Val/Val patients having greaterfrontal lobe activation than Met/Met patients whileperforming the one-back working memory task isconsistent with Egan et al study.52 The authors foundgreater fMRI BOLD response in the DLPFC andanterior cingulate among unaffected siblings of

COMT and frontal lobe functions and morphology in schizophreniaB-C Ho et al

294

Molecular Psychiatry

schizophrenia patients with Val/Val genotype. Sinceour Val/Val patients had comparable reaction times asour Met/Met patients and the one-back task is a lowload working memory test, higher frontal lobe rCBFamong our Val/Val patients suggests that they may be‘working harder’ in order to achieve equal taskperformance. While relative prefrontal cortical ineffi-ciency in the valine allele may be a plausibleinterpretation of these PET study findings, a diame-trically opposite view could be that lower frontal loberCBF among Met/Met patients may represent hypo-frontality relative to Val/Val patients. Both hypofron-tality as well as hyperfrontality have been describedin schizophrenia. Unfortunately, we did not havesufficient numbers of healthy volunteers with Met/Met and Val/Val genotypes to assess the patients’relative brain activations. As Manoach117 has ele-gantly presented, the interpretation of functionalneuroimaging findings in schizophrenia is complex,and a variety of reasons (ranging from methodologicalissues to heterogeneity of schizophrenia) may accountfor the seemingly discrepant findings of both hypo-frontality as well as increased prefrontal activity inschizophrenia.

Irrespective of how these PET findings may beinterpreted, that Val/Val patients and Met/Met pa-tients differed significantly on frontal lobe rCBF seeminconsistent with the rest of our findings of nosignificant COMT genotype effects on a variety offrontal lobe-related measures. Cerebral blood flow,being a more direct measure of neuronal function,may be more sensitive for assessing the effects ofCOMT genotype than are neuropsychological mea-sures or MRI brain morphometry. Conversely, our PETstudy involved only 13 patients, and the largersample size used in the rest of this study may provideresults that can be better generalized to the averagepatient with schizophrenia. Future efforts to integratefunctional neuroimaging toward better understandingthe effects of COMT genotype on cognitive endophe-notypes of schizophrenia will require large samples ofboth patients as well as healthy volunteers. Workingmemory load will also need to be varied in futurefunctional neuroimaging studies so as to be better ableto interpret differences in brain activation betweenCOMT genotypes.

Conclusion

The complex genetics and phenotypic heterogeneityremain as major obstacles in schizophrenia research.Integrating multimodal investigational approacheswill help address these challenges, and facilitategreater understanding of the neurobiological basis forthis group of disorders.

Acknowledgements

This research was supported in part by NIMH GrantsMH31593, MH40856 and MH43271. Parts of thisresearch were presented at the IXth International

Congress on Schizophrenia Research, ColoradoSprings, CO, March 28–April 2, 2003.

References

1 Andreasen NC. Schizophrenia: the fundamental questions. BrainRes Brain Res Rev 2000; 31: 106–112.

2 Kennedy JL, Farrer LA, Andreasen NC, Mayeux R, St George-Hyslop P. The genetics of adult-onset neuropsychiatric disease:complexities and conundra? Science 2003; 302: 822–826.

3 Moldin SO. Indicators of liability to schizophrenia: perspectivesfrom genetic epidemiology. Schizophr Bull 1994; 20: 169–184.

4 Thaker GK. Current progress in schizophrenia research. Searchfor genes of schizophrenia: back to defining valid phenes. J NervMent Dis 2002; 190: 411–412.

5 Almasy L, Blangero J. Endophenotypes as quantitative risk factorsfor psychiatric disease: rationale and study design. Am J MedGenet 2001; 105: 42–44.

6 Tsuang MT, Faraone SV. The frustrating search for schizophreniagenes. Am J Med Genet (Semin Med Genet) 2000; 97: 1–3.

7 Gottesman II, Gould TD. The endophenotype concept inpsychiatry: etymology and strategic intentions. Am J Psychiatry2003; 160: 636–645.

8 Arolt V, Lencer R, Nolte A, Muller-Myhsok B, Purmann S,Schurmann M et al. Eye tracking dysfunction is a putativephenotypic susceptibility marker of schizophrenia and maps to alocus on chromosome 6p in families with multiple occurrence ofthe disease. Am J Med Genet 1996; 67: 564–579.

9 Freedman R, Coon H, Myles-Worsley M, Orr-Urtreger A, OlincyA, Davis A et al. Linkage of a neurophysiological deficit inschizophrenia to a chromosome 15 locus. Proc Natl Acad SciUSA 1997; 94: 587–592.

10 Liu H, Heath SC, Sobin C, Roos JL, Galke BL, Blundell ML et al.Genetic variation at the 22q11 PRODH2/DGCR6 locus presents anunusual pattern and increases susceptibility to schizophrenia.Proc Natl Acad Sci USA 2002; 99: 3717–3722.

11 Leal SM. Phenotypes and genetic analysis of psychiatric andneuropsychiatric traits. Am J Med Genet 2001; 105: 4–7.

12 Baddeley A. Working memory. Science 1992; 255: 556–559.13 Cohen JD, Braver TS, O’Reilly RC. A computational approach to

prefrontal cortex, cognitive control and schizophrenia: recentdevelopments and current challenges. Philos Trans R Soc LondSer B 1996; 351: 1515–1527.

14 Goldman-Rakic PS. Prefrontal cortical dysfunction in schizo-phrenia: the relevance of working memory. In: Carroll BJ, BarrettJE (eds). Psychopathology and the Brain. Raven Press: New York,NY, 1990 pp 1–23.

15 Park S, Holzman PS. Schizophrenics show spatial workingmemory deficits. Arch Gen Psychiatry 1992; 49: 975–982.

16 Park S, Holzman PS, Goldman-Rakic PS. Spatial workingmemory deficits in the relatives of schizophrenic patients. ArchGen Psychiatry 1995; 52: 821–828.

17 Goldberg TE, Torrey EF, Gold JM, Ragland JD, Bigelow LB,Weinberger DR. Learning and memory in monozygotic twinsdiscordant for schizophrenia. Psychol Med 1993; 23: 71–85.

18 Cannon TD, Bearden CE, Hollister JM, Rosso IM, Sanchez LE,Hadley T. Childhood cognitive functioning in schizophreniapatients and their unaffected siblings: a prospective cohort study.Schizophr Bull 2000; 26: 379–393.

19 Myles-Worsley M, Park S. Spatial working memory deficits inschizophrenia patients and their first degree relatives from Palau,Micronesia. Am J Med Genet 2002; 114: 609–615.

20 Tuulio-Henriksson A, Arajarvi R, Partonen T, Haukka J, Varilo T,Schreck M et al. Familial loading associates with impairment invisual span among healthy siblings of schizophrenia patients.Biol Psychiatry 2003; 54: 623–628.

21 Tuulio-Henriksson A, Haukka J, Partonen T, Varilo T, Paunio T,Ekelund J et al. Heritability and number of quantitative trait lociof neurocognitive functions in families with schizophrenia. Am JMed Genet 2002; 114: 483–490.

22 Conklin HM, Curtis CE, Katsanis J, Iacono WG. Verbalworking memory impairment in schizophrenia patients and their

COMT and frontal lobe functions and morphology in schizophreniaB-C Ho et al

295

Molecular Psychiatry

first-degree relatives: evidence from the digit span task. Am JPsychiatry 2000; 157: 275–277.

23 Farmer CM, O’Donnell BF, Niznikiewicz MA, Voglmaier MM,McCarley RW, Shenton ME. Visual perception and workingmemory in schizotypal personality disorder. Am J Psychiatry2000; 157: 781–788.

24 Ando J, Ono Y, Wright MJ. Genetic structure of spatial and verbalworking memory. Behav Genet 2001; 31: 615–624.

25 Cannon TD, Huttunen MO, Lonnqvist J, Tuulio-Henriksson A,Pirkola T, Glahn D et al. The inheritance of neuropsychologicaldysfunction in twins discordant for schizophrenia. Am J HumGenet 2000; 67: 369–382.

26 Gold S, Arndt S, Nopoulos P, O’Leary DS, Andreasen NC.Longitudinal study of cognitive function in first-episodeand recent-onset schizophrenia. Am J Psychiatry 1999; 156:1342–1348.

27 Goldberg TE, Weinberger DR. Effects of neuroleptic medicationson the cognition of patients with schizophrenia: a review ofrecent studies. J Clin Psychiatry 1996; 57: 62–65.

28 Carter C, Robertson L, Nordahl T, Chaderjian M, Kraft L, O’Shora-Celaya L. Spatial working memory deficits and their relationshipto negative symptoms in unmedicated schizophrenia patients.Biol Psychiatry 1996; 40: 930–932.

29 Barch DM, Carter CS, Braver TS, Sabb FW, MacDonald III A, NollDC et al. Selective deficits in prefrontal cortex function inmedication-naive patients with schizophrenia. Arch Gen Psy-chiatry 2001; 58: 280–288.

30 Keefe RS, Silva SG, Perkins DO, Lieberman JA. The effects ofatypical antipsychotic drugs on neurocognitive impairment inschizophrenia: a review and meta-analysis. Schizophr Bull 1999;25: 201–222.

31 Friedman HR, Goldman-Rakic PS. Coactivation of prefrontalcortex and inferior parietal cortex in working memory tasksrevealed by 2DG functional mapping in the rhesus monkey.J Neurosci 1994; 14(Part 1): 2775–2788.

32 Petrides M, Alivisatos B, Meyer E, Evans AC. Functionalactivation of the human frontal cortex during the performanceof verbal working memory tasks. Proc Natl Acad Sci USA 1993;90: 878–882.

33 Lewis DA, Campbell MJ, Foote SL, Goldstein M, Morrison JH.The distribution of tyrosine hydroxylase-immunoreactive fibersin primate neocortex is widespread but regionally specific.J Neurosci 1987; 7: 279–290.

34 Levitt P, Rakic P, Goldman-Rakic P. Region-specific distributionof catecholamine afferents in primate cerebral cortex: afluorescence histochemical analysis. J Comp Neurol 1984; 227:23–36.

35 Lidow MS, Goldman-Rakic PS, Gallager DW, Rakic P. Distributionof dopaminergic receptors in the primate cerebral cortex:quantitative autoradiographic analysis using [3H]raclopride,[3H]spiperone and [3H]SCH23390. Neuroscience 1991; 40:657–671.

36 Hurd YL, Suzuki M, Sedvall GC. D1 and D2 dopamine receptormRNA expression in whole hemisphere sections of the humanbrain. J Chem Neuroanat 2001; 22: 127–137.

37 Brozoski TJ, Brown RM, Rosvold HE, Goldman PS. Cognitivedeficit caused by regional depletion of dopamine in prefrontalcortex of rhesus monkey. Science 1979; 205: 929–932.

38 Sawaguchi T, Goldman-Rakic PS. The role of D1-dopaminereceptor in working memory: local injections of dopamineantagonists into the prefrontal cortex of rhesus monkeysperforming an oculomotor delayed-response task. J Neurophysiol1994; 71: 515–528.

39 Callicott JH, Bertolino A, Mattay VS, Langheim FJ, Duyn J,Coppola R et al. Physiological dysfunction of the dorsolateralprefrontal cortex in schizophrenia revisited. Cereb Cortex 2000;10: 1078–1092.

40 Stevens AA, Goldman-Rakic PS, Gore JC, Fulbright RK, WexlerBE. Cortical dysfunction in schizophrenia during auditory wordand tone working memory demonstrated by functional magneticresonance imaging. Arch Gen Psychiatry 1998; 55: 1097–1103.

41 Honey GD, Bullmore ET, Soni W, Varatheesan M, Williams SC,Sharma T. Differences in frontal cortical activation by a workingmemory task after substitution of risperidone for typical

antipsychotic drugs in patients with schizophrenia. Proc NatlAcad Sci USA 1999; 96: 13432–13437.

42 Manoach DS, Press DZ, Thangaraj V, Searl MM, Goff DC, HalpernE et al. Schizophrenic subjects activate dorsolateral prefrontalcortex during a working memory task, as measured by fMRI. BiolPsychiatry 1999; 45: 1128–1137.

43 Manoach DS, Gollub RL, Benson ES, Searl MM, Goff DC, HalpernE et al. Schizophrenic subjects show aberrant fMRI activation ofdorsolateral prefrontal cortex and basal ganglia during workingmemory performance. Biol Psychiatry 2000; 48: 99–109.

44 Karoum F, Chrapusta SJ, Egan MF. 3-Methoxytyramine is themajor metabolite of released dopamine in the rat frontal cortex:reassessment of the effects of antipsychotics on the dynamicsof dopamine release and metabolism in the frontal cortex,nucleus accumbens, and striatum by a simple two pool model.J Neurochem 1994; 63: 972–979.

45 Huotari M, Gogos JA, Karayiorgou M, Koponen O, Forsberg M,Raasmaja A et al. Brain catecholamine metabolism in catechol-O-methyltransferase (COMT)-deficient mice. Eur J Neurosci 2002;15: 246–256.

46 Wayment HK, Schenk JO, Sorg BA. Characterization of extra-cellular dopamine clearance in the medial prefrontal cortex:role of monoamine uptake and monoamine oxidase inhibition.J Neurosci 2001; 21: 35–44.

47 Moron JA, Brockington A, Wise RA, Rocha BA, Hope BT.Dopamine uptake through the norepinephrine transporter inbrain regions with low levels of the dopamine transporter:evidence from knock-out mouse lines. J Neurosci 2002; 22:389–395.

48 Palmatier MA, Kang AM, Kidd KK. Global variation in thefrequencies of functionally different catechol-O-methyltransfer-ase alleles. Biol Psychiatry 1999; 46: 557–567.

49 Lotta T, Vidgren J, Tilgmann C, Ulmanen I, Melen K, Julkunen Iet al. Kinetics of human soluble and membrane-bound catecholO-methyltransferase: a revised mechanism and description ofthe thermolabile variant of the enzyme. Biochemistry 1995; 34:4202–4210.

50 Badner JA, Gershon ES. Meta-analysis of whole-genome linkagescans of bipolar disorder and schizophrenia. Mol Psychiatry2002; 7: 405–411.

51 Murphy KC, Jones LA, Owen MJ. High rates of schizophrenia inadults with velo-cardio-facial syndrome. Arch Gen Psychiatry1999; 56: 940–945.

52 Egan MF, Goldberg TE, Kolachana BS, Callicott JH, Mazzanti CM,Straub RE et al. Effect of COMT Val108/158 Met genotype onfrontal lobe function and risk for schizophrenia. Proc Natl AcadSci USA 2001; 98: 6917–6922.

53 Weinberger DR, Egan MF, Bertolino A, Callicott JH, Mattay VS,Lipska BK et al. Prefrontal neurons and the genetics ofschizophrenia. Biol Psychiatry 2001; 50: 825–844.

54 Daniels JK, Williams NM, Williams J, Jones LA, Cardno AG,Murphy KC et al. No evidence for allelic association betweenschizophrenia and a polymorphism determining high or lowcatechol O-methyltransferase activity. Am J Psychiatry 1996; 153:268–270.

55 Karayiorgou M, Gogos JA, Galke BL, Wolyniec PS, Nestadt G,Antonarakis SE et al. Identification of sequence variants andanalysis of the role of the catechol-O-methyl-transferase gene inschizophrenia susceptibility. Biol Psychiatry 1998; 43: 425–431.

56 Li T, Sham PC, Vallada H, Xie T, Tang X, Murray RM et al.Preferential transmission of the high activity allele of COMT inschizophrenia. Psychiatr Genet 1996; 6: 131–133.

57 Kunugi H, Vallada HP, Sham PC, Hoda F, Arranz MJ, Li T et al.Catechol-O-methyltransferase polymorphisms and schizophre-nia: a transmission disequilibrium study in multiplying affectedfamilies. Psychiatr Genet 1997; 7: 97–101.

58 de Chaldee M, Laurent C, Thibaut F, Martinez M, Samolyk D,Petit M et al. Linkage disequilibrium on the COMT gene inFrench schizophrenics and controls. Am J Med Genet 1999; 88:452–457.

59 Joober R, Gauthier J, Lal S, Bloom D, Lalonde P, Rouleau G et al.Catechol-O-methyltransferase Val-108/158-Met gene variants as-sociated with performance on the Wisconsin Card Sorting Test.Arch Gen Psychiatry 2002; 59: 662–663.

COMT and frontal lobe functions and morphology in schizophreniaB-C Ho et al

296

Molecular Psychiatry

60 Park TW, Yoon KS, Kim JH, Park WY, Hirvonen A, Kang D.Functional catechol-O-methyltransferase gene polymorphismand susceptibility to schizophrenia. Eur Neuropsychopharmacol2002; 12: 299–303.

61 Norton N, Kirov G, Zammit S, Jones G, Jones S, Owen R et al.Schizophrenia and functional polymorphisms in the MAOA andCOMT genes: no evidence for association or epistasis. Am J MedGenet 2002; 114: 491–496.

62 Glatt SJ, Faraone SV, Tsuang MT. Association between afunctional catechol O-methyltransferase gene polymorphismand schizophrenia: meta-analysis of case–control and family-based studies. Am J Psychiatry 2003; 160: 469–476.

63 Wonodi I, Stine OC, Mitchell BD, Buchanan RW, Thaker GK.Association between Val108/158 Met polymorphism of theCOMT gene and schizophrenia. Am J Med Genet 2003; 120B:47–50.

64 Andreasen NC, Flaum M, Arndt S. The Comprehensive Assess-ment of Symptoms and History (CASH). An instrument forassessing diagnosis and psychopathology. Arch Gen Psychiatry1992; 49: 615–623.

65 Lahiri DK, Nurnberger Jr JI. A rapid non-enzymatic method forthe preparation of HMW DNA from blood for RFLP studies.Nucleic Acids Res 1991; 19: 5444.

66 Perry W, Heaton RK, Potterat E, Roebuck T, Minassian A, BraffDL. Working memory in schizophrenia: transient ‘online’ storageversus executive functioning. Schizophr Bull 2001; 27: 157–176.

67 Pukrop R, Matuschek E, Ruhrmann S, Brockhaus-Dumke A,Tendolkar I, Bertsch A et al. Dimensions of working memorydysfunction in schizophrenia. Schizophr Res 2003; 62:259–268.

68 Heaton RK, Chelune GK, Talley JL, Kay GG, Curtiss G. WisconsinCard Sorting Test Manual. Psychological Assessment ResourcesInc.: Odessa, FL, 1993.

69 Reitan RM. Validity of the Trail Making Test as an indicator oforganic brain damage. Percept Motor Skills 1958; 8: 271–276.

70 Jarvis PE, Barth JT. Halstead-Reitan Test Battery: An Inter-pretative Guide. Psychological Assessment Resources: Odessa,FL, 1986.

71 Andreasen NC, Cohen G, Harris G, Cizadlo T, Parkkinen J, RezaiK et al. Image processing for the study of brain structure andfunction: problems and programs. J Neuropsychiatry ClinNeurosci 1992; 4: 125–133.

72 Andreasen NC, Cizadlo T, Harris G, Swayze V, O’Leary DS, CohenG et al. Voxel processing techniques for the antemortem study ofneuroanatomy and neuropathology using magnetic resonanceimaging. J Neuropsychiatry Clin Neurosci 1993; 5: 121–130.

73 Andreasen NC, Harris G, Cizadlo T, Arndt S, O’Leary DS, SwayzeV et al. Techniques for measuring sulcal/gyral patterns in thebrain as visualized through magnetic resonance scanning:BRAINPLOT and BRAINMAP. Proc Natl Acad Sci USA 1994;91: 93–97.

74 Magnotta VA, Heckel D, Andreasen NC, Cizadlo T, Corson PW,Ehrhardt JC et al. Measurement of brain structures with artificialneural networks: two- and three-dimensional applications.Radiology 1999; 211: 781–790.

75 Woods RP, Cherry SR, Mazziotta JC. Rapid automated algorithmfor aligning and reslicing PET images. J Comput Assist Tomogr1992; 16: 620–633.

76 Talairach J, Tournoux P. Co-Planar Stereotaxic Atlas of theHuman Brain. Thieme Medical Publishers: New York, 1988.

77 Andreasen NC, Rajarethinam R, Cizadlo T, Arndt S, Swayze IIVW, Flashman LA et al. Automatic atlas-based volume estimationof human brain regions from MR images. J Comput Assist Tomogr1996; 20: 98–106.

78 Harris G, Andreasen NC, Cizadlo T, Bailey JM, Bockholt HJ,Magnotta VA et al. Improving tissue classification in MRI: athree-dimensional multispectral discriminant analysis methodwith automated training class selection. J Comput Assist Tomogr1999; 23: 144–154.

79 Herscovitch P, Markham J, Raichle ME. Brain blood flowmeasured with intravenous H2(15)O. I. Theory and erroranalysis. J Nucl Med 1983; 24: 782–789.

80 Hichwa RD, Boles Ponto LL, Watkins GL. Clinical blood flowmeasurement with [15O]water and positron emission tomography

(PET). In: Emran AM (ed). Chemists’ Views of Imaging Centers.Plenum Press: New York, 1995 pp 401–417.

81 Arndt S, Cizadlo T, Andreasen NC, Heckel D, Gold S, O’Leary DS.Tests for comparing images based on randomization andpermutation methods. J Cereb Blood Flow Metab 1996; 16:1271–1279.

82 Strous RD, Bark N, Woerner M, Lachman HM. Lack of associationof a functional catechol-O-methyltransferase gene polymorphismin schizophrenia. Biol Psychiatry 1997; 41: 493–495.

83 Chen CH, Lee YR, Wei FC, Koong FJ, Hwu HG, Hsiao KJ.Association study of NlaIII and MspI genetic polymorphisms ofcatechol-O-methyltransferase gene and susceptibility to schizo-phrenia. Biol Psychiatry 1997; 41: 985–987.

84 Chen CH, Lee YR, Chung MY, Wei FC, Koong FJ, Shaw CK et al.Systematic mutation analysis of the catechol O-methyltransferasegene as a candidate gene for schizophrenia. Am J Psychiatry 1999;156: 1273–1275.

85 Arinami T, Ohtsuki T, Takase K, Shimizu H, Yoshikawa T,Horigome H et al. Screening for 22q11 deletions in aschizophrenia population. Schizophr Res 2001; 52:167–170.

86 Herken H, Erdal ME. Catechol-O-methyltransferase genepolymorphism in schizophrenia: evidence for associationbetween symptomatology and prognosis. Psychiatr Genet 2001;11: 105–109.

87 Liou YJ, Tsai SJ, Hong CJ, Wang YC, Lai IC. Association analysisof a functional catechol-o-methyltransferase gene polymorphismin schizophrenic patients in Taiwan. Neuropsychobiology 2001;43: 11–14.

88 Inada T, Nakamura A, Iijima Y. Relationship between catechol-O-methyltransferase polymorphism and treatment-resistant schizo-phrenia. Am J Med Genet 2003; 120B: 35–39.

89 Shifman S, Bronstein M, Sternfeld M, Pisante-Shalom A, Lev-Lehman E, Weizman A et al. A highly significant associationbetween a COMT haplotype and schizophrenia. Am J Hum Genet2002; 71: 1296–1302.

90 Ohmori O, Shinkai T, Kojima H, Terao T, Suzuki T, Mita T et al.Association study of a functional catechol-O-methyltransferasegene polymorphism in Japanese schizophrenics. Neurosci Lett1998; 243: 109–112.

91 Kotler M, Barak P, Cohen H, Averbuch IE, Grinshpoon A,Gritsenko I et al. Homicidal behavior in schizophrenia associatedwith a genetic polymorphism determining low catechol O-methyltransferase (COMT) activity. Am J Med Genet 1999; 88:628–633.

92 Wei J, Hemmings GP. Lack of evidence for association betweenthe COMT locus and schizophrenia. Psychiatr Genet 1999; 9:183–186.

93 de Chaldee M, Corbex M, Campion D, Jay M, Samolyk D, Petit Met al. No evidence for linkage between COMT and schizophreniain a French population. Psychiatry Res 2001; 102: 87–90.

94 Semwal P, Prasad S, Bhatia T, Deshpande SN, Wood J,Nimgaonkar VL et al. Family-based association studies ofmonoaminergic gene polymorphisms among North Indians withschizophrenia. Mol Psychiatry 2001; 6: 220–224.

95 Lohmueller KE, Pearce CL, Pike M, Lander ES, Hirschhorn JN.Meta-analysis of genetic association studies supports a contribu-tion of common variants to susceptibility to common disease. NatGenet 2003; 33: 177–182.

96 Strous RD, Bark N, Parsia SS, Volavka J, Lachman HM. Analysisof a functional catechol-O-methyltransferase gene polymorphismin schizophrenia: evidence for association with aggressive andantisocial behavior. Psychiatry Res 1997; 69: 71–77.

97 Lachman HM, Nolan KA, Mohr P, Saito T, Volavka J. Associationbetween catechol O-methyltransferase genotype and violence inschizophrenia and schizoaffective disorder. Am J Psychiatry1998; 155: 835–837.

98 Nolan KA, Volavka J, Czobor P, Cseh A, Lachman H, Saito T et al.Suicidal behavior in patients with schizophrenia is related toCOMT polymorphism. Psychiatr Genet 2000; 10: 117–124.

99 Jones G, Zammit S, Norton N, Hamshere ML, Jones SJ, Milham Cet al. Aggressive behaviour in patients with schizophrenia isassociated with catechol-O-methyltransferase genotype. Br JPsychiatry 2001; 179: 351–355.

COMT and frontal lobe functions and morphology in schizophreniaB-C Ho et al

297

Molecular Psychiatry

100 Illi A, Kampman O, Anttila S, Roivas M, Mattila KM, Lehtimaki Tet al. Interaction between angiotensin-converting enzyme andcatechol-O-methyltransferase genotypes in schizophrenics withpoor response to conventional neuroleptics. Eur Neuropsycho-pharmacol 2003; 13: 147–151.

101 Malhotra AK, Kestler LJ, Mazzanti C, Bates JA, Goldberg T,Goldman D. A functional polymorphism in the COMT gene andperformance on a test of prefrontal cognition. Am J Psychiatry2002; 159: 652–654.

102 Goldberg TE, Egan MF, Gscheidle T, Coppola R, Weickert T,Kolachana BS et al. Executive subprocesses in working memory:relationship to catechol-O-methyltransferase Val158Metgenotype and schizophrenia. Arch Gen Psychiatry 2003; 60:889–896.

103 Gallinat J, Bajbouj M, Sander T, Schlattmann P, Xu K, Ferro EF etal. Association of the G1947A COMT (Val(108/158)Met) genepolymorphism with prefrontal P300 during information proces-sing. Biol Psychiatry 2003; 54: 40–48.

104 Tsai SJ, Yu YW, Chen TJ, Chen JY, Liou YJ, Chen MC et al.Association study of a functional catechol-O-methyltransferase-gene polymorphism and cognitive function in healthy females.Neurosci Lett 2003; 338: 123–126.

105 Bilder RM, Volavka J, Czobor P, Malhotra AK, Kennedy JL, Ni X etal. Neurocognitive correlates of the COMT Val(158)Metpolymorphism in chronic schizophrenia. Biol Psychiatry 2002;52: 701–707.

106 Nolan KA, Bilder RM, Lachman HM, Volavka J. Catechol O-methyltransferase Val158Met polymorphism in schizophrenia:differential effects of Val and Met alleles on cognitive stabilityand flexibility. Am J Psychiatry 2004; 161: 359–361.

107 Fossella J, Sommer T, Fan J, Wu Y, Swanson JM, Pfaff DW et al.Assessing the molecular genetics of attention networks. BMCNeurosci 2002; 3: 14.

108 Joober R, Rouleau GA, Lal S, Dixon M, O’Driscoll G, Palmour R etal. Neuropsychological impairments in neuroleptic-responder vsnonresponder schizophrenic patients and healthy volunteers.Schizophr Res 2002; 53: 229–238.

109 Rinne JO. Muscarinic and dopaminergic receptors in the aginghuman brain. Brain Res 1987; 404: 162–168.

110 de Keyser J, De Backer JP, Vauquelin G, Ebinger G. The effect ofaging on the D1 dopamine receptors in human frontal cortex.Brain Res 1990; 528: 308–310.

111 Suhara T, Fukuda H, Inoue O, Itoh T, Suzuki K, Yamasaki T et al.Age-related changes in human D1 dopamine receptors measuredby positron emission tomography. Psychopharmacology (Berl)1991; 103: 41–45.

112 Kaasinen V, Vilkman H, Hietala J, Nagren K, Helenius H,Olsson H et al. Age-related dopamine D2/D3 receptor loss inextrastriatal regions of the human brain. Neuobiol Aging 2000;21: 683–688.

113 Volkow ND, Logan J, Fowler JS, Wang GJ, Gur RC, Wong C et al.Association between age-related decline in brain dopamineactivity and impairment in frontal and cingulate metabolism.Am J Psychiatry 2000; 157: 75–80.

114 Goldman-Rakic PS, Brown RM. Regional changes of monoaminesin cerebral cortex and subcortical structures of aging rhesusmonkeys. Neuroscience 1981; 6: 177–187.

115 Wassink TH, Crowe RR, Andreasen NC. Tumor necrosis factorreceptor—II: heritability and effect on brain morphology inschizophrenia. Mol Psychiatry 2000; 5: 678–682.

116 Wassink TH, Nopoulos P, Pietila J, Crowe RR, Andreasen NC.NOTCH4 and the frontal lobe in schizophrenia. Am J Med Genet2003; 118B: 1–7.

117 Manoach DS. Prefrontal cortex dysfunction during workingmemory performance in schizophrenia: reconciling discrepantfindings. Schizophr Res 2003; 60: 285–298.

COMT and frontal lobe functions and morphology in schizophreniaB-C Ho et al

298

Molecular Psychiatry