COMT genotype predicts BOLD signal and noise characteristics in prefrontal circuits
Introduction
In 2001, Egan et al. reported that a single nucleotide polymorphism (Val108/158Met) in the catchol-O-methyltransferase (COMT) gene predicts executive cognition and the efficiency of prefrontal function in humans as assayed with functional magnetic resonance imaging (fMRI) during a working memory task. Since this initial report, several neuropsychological, electrophysiological and functional neuroimaging studies have been published which essentially confirmed and extended the original findings (Bilder et al., 2002, Malhotra et al., 2002, Mattay et al., 2003, Gallinat et al., 2003, Enoch et al., 2003, Goldberg et al., 2003, Diamond et al., 2004, Rosa et al., 2004, Foltynie et al., 2004, Bertolino et al., 2004, Weickert et al., 2004, Bearden et al., 2004, Blasi et al., 2004, Meyer-Lindenberg et al., 2005, De Frias et al., 2005, Bruder et al., 2005, Galderisi et al., 2005, Baker et al., 2005, Winterer et al., 2006). In line with these results are studies of COMT knockout mice (Gogos et al., 1998) as well as pharmacological investigations with COMT inhibitors showing specific prefrontal effects on dopamine (DA) levels (Turnbridge et al., 2004) and improved cognitive performance both in rats (Khromova et al., 1997, Liljequist et al., 1997) and in humans (Gasparini et al., 1997, Apud et al., in press). The particular scientific meaning of this work is that the COMT Val108/158Met single nucleotide polymorphism (SNP) possibly contributes to the risk for schizophrenia illness (Glatt et al., 2003, Harrison and Weinberger, 2005, Williams et al., 2005) and that it is the first SNP, which has been consistently associated with human prefrontal function (Weinberger et al., 2001, Winterer and Goldman, 2003). In addition, this work is contributing to our knowledge about the neuromodulatory action of DA on prefrontal brain function, which is important because there is still a lack of clear understanding of the basic principles of actions of DA in the prefrontal cortex (Seamans and Yang, 2003).
The COMT gene codes for the dopamine DA catabolizing enzyme COMT and contains a functional and common variation in its coding sequence, i.e., a substitution of valine (Val) by methionine (Met) in the peptide sequence, which is caused by a transition of guanine to adenine at codon 158 (Lotta et al., 1995, Lachman et al., 1996). This single amino acid substitution affects the temperature lability of the enzyme; at body temperature, the Met allele has less than half of the enzyme activity as compared to the Val allele (Lotta et al., 1995, Lachman et al., 1996, Weinshilboum et al., 1999, Chen et al., 2004). These data suggest that individuals with Val alleles would have relatively greater inactivation of prefrontal DA and therefore less effective prefrontal DA signaling which in turn provides an explanation for its impact on cognition since DA agonists are well known to affect physiologic correlates of human frontal-lobe-related cognitive function (Mattay et al., 1996).
While there is consistent evidence that the COMT Val108/158Met SNP impacts on frontal-lobe-related cognitive function, the precise physiological mechanism by which this is achieved is not clear. Based on electrophysiological studies during attention-requiring and working memory tasks with DA agonists in non-human primates (Sawaguchi et al., 1986, Sawaguchi et al., 1990, Williams and Goldman-Rakic, 1995) as well as functional imaging studies of small animals with dopaminergic drugs (Chen et al., 1997), one would expect that Val carriers with less available synaptic DA would show a diminished signal-to-noise ratio (SNR) of prefrontal activity during working memory or attention-requiring tasks. In fact, when discussing their data on the effects of the COMT gene variation on prefrontal function, the authors of the abovementioned imaging studies frequently emphasize that their findings may point to a lower neuronal SNR in Val carriers. However, this interpretation is not easily reconciled with their data. This is because all previously conducted functional magnetic resonance imaging (fMRI) studies found the opposite in Val carriers, i.e., an enhanced BOLD (blood-oxygenation-level-dependent) signal in dorsolateral prefrontal and anterior cingulate cortex either during a working memory task (Egan et al., 2001, Mattay et al., 2003, Bertolino et al., 2004) or during an attentional control task (Blasi et al., 2004). In both experimental conditions, a parametric task design was chosen, that is, the level of task difficulty was systematically varied whereby the observed COMT genotype effect on the prefrontal BOLD response increased with cognitive load consistent with recent neuropsychological studies showing a linear increase of COMT genotype effect on working memory performance or attention as a function of task difficulty (Goldberg et al., 2003, Blasi et al., 2004). When interpreting their results, the authors generally point out that the observation of an enhanced BOLD response in Val carriers may reflect a compensatory, “less efficient” task-related brain response resulting from a diminished neuronal SNR, more or less explicitly referring to the frequently reported observations that more difficult tasks are associated with increased brain activation (Grasby et al., 1994) while task practice is correlated with decreased brain activation (Buckner et al., 1998). This interpretation is also considered to be consistent with evidence that patients with Parkinson's disease show less efficient cortical processing during working memory tasks when they are DA depleted than when they are DA repleted (Mattay et al., 2002, Cools et al., 2002).
This interpretation, however, does not take into account recent comparative studies of the BOLD signal and intracortical electrophysiological recordings as conducted, for instance, by Logothetis et al. (2001). Their findings rather suggest a positive correlation between neuronal SNR and BOLD response. On the other hand, it is also conceivable that the maximal BOLD response in the peak voxels may be greater in Met carriers while in Val carriers the BOLD response is spatially more extended. Given these difficulties when interpreting COMT genotype effects on the BOLD response, scalp-recorded electrophysiological studies in humans, although spatially less accurate, might be helpful for interpretation since they provide a direct measure of neuronal activity. So far, resting EEG (electroencephalogram) and event-related potentials studies reported findings that are more in line with the expectation from animal data of a DA-induced enhancement of neuronal SNR (s.a.). Enoch et al. (2003) reported that Val carriers are characterized by an increase of background noise (alpha-power) in resting EEG. Furthermore, Baker et al. (2005) found that Val carriers have a lower prefrontal signal amplitude of the mismatch negativity event-related potential. In contrast, Gallinat et al. (2003) found that Val carriers show a higher frontal P300 signal amplitude. It has been argued, however, that the frontal P300 characteristically shows a very low amplitude and that the finding of an enhanced frontal P300 amplitude in Val carriers could reflect that event-related electrophysiological noise, i.e., event-related response variability, is increased (Winterer et al., 2003, Winterer et al., 2004). In fact, a follow-up event-related P300 potential study in an independent sample, which directly addressed the question of whether prefrontal electrophysiological noise is increased in Val carriers, found that this is the case (Winterer et al., 2006). Together, these electrophysiological studies indicate that Val carriers show apparently higher electrophysiological noise levels while the signal amplitude is decreased, i.e., the SNR is reduced—a finding which was recently further substantiated by Stefanis et al. (2005) who found that reaction variability during an attention-requiring Continuous Performance Task is increased in Val carriers. In light of these electrophysiological and behavioral findings, the question arises whether the previously conducted fMRI studies of an increased task-related BOLD signal in Val carriers reflected increased noise primarily in the spatiotemporal domain. In other words, while Met carriers may show a focused BOLD response in a circumscribed area, Val carriers would be characterized by a spatially extended “noisy” BOLD response pattern. This interpretation of the data gains some support from earlier PET (positron emission tomography) studies, which tested the effect of DA agonists on cerebral blood flow (Weinberger et al., 1988, Daniel et al., 1989, Daniel et al., 1991, Mattay et al., 1996). In these studies, the authors interpreted their data as reflecting a sharpening of the signal, i.e., increase SNR. Unfortunately, these earlier studies did not quantify this possible effect of DA agonists but merely conveyed their subjective impression.
In the present study, we addressed the question of whether the prefrontal SNR of the BOLD response is decreased in Val carriers using a visual oddball task and an approach to analysis of fMRI data that maximizes noise characterization (Winterer et al., in press, Musso et al., 2006). In previous work, we have shown that prefrontal activation with maximum activation in the supplementary motor cortex (SMA) during comparable task conditions is related to ‘motivation’ respectively ‘task engagement’ and ‘selective attention’ (Winterer et al., 2001, Winterer et al., 2002, Winterer et al., in press, Mulert et al., 2001, Gallinat et al., 2002, Musso et al., 2006). Analyses of fMRI indices were done separately for the region of the peak BOLD response in SMA as well as for the surrounding medial frontal lobe (MFL).
Section snippets
Subjects
N = 44 unrelated healthy Caucasian (European) subjects (20 males, age: 22.7 ± 1.7 years, all right-handed) were investigated with fMRI and genotyped for COMT Val/Met. Participants were only included if there was no evidence of a medical or neurological condition that could interfere with the purpose of the study or if there was a history for any psychiatric DSM-IV axis I or axis II disorder including current or recent drug or alcohol abuse as assessed by a Structured Clinical Interview (First et
Experimental and subject parameters
Task and subject parameters are provided in Table 1. Genotype distribution with a relatively high prevalence of Met carriers in our sample of European Caucasians is in accordance with a previous independent investigation conducted in Germany (Gallinat et al., 2003). The only significant between-genotype group difference was found for age, which, however, was minimal. Genotype had no statistically impact on behavioral performance. Even so, Met carriers showed numerically faster and less variable
Discussion
Earlier electrophysiological studies in non-human primates have shown that the modulatory role of DA on neuronal activity critically determines the SNR of neuronal firing in prefrontal cortex (Sawaguchi et al., 1986, Sawaguchi et al., 1990, Williams and Goldman-Rakic, 1995). Using COMT genotype as a probe, Winterer et al. (2006) demonstrated this DA effect in humans by means of scalp-recorded electrophysiological recordings, that is, a higher prefrontal level of event-related electromagnetic
References (80)
- et al.
COMT Val108/158 Met modifies mismatch negativity and cognitive function in 22q11 deletion syndrome
Biol. Psychiatry
(2005) - et al.
Neurocognitive correlates of the COMT Val(158)Met polymorphism in chronic schizophrenia
Biol. Psychiatry
(2002) - et al.
Catechol-O-methyltransferase (COMT) genotypes and working memory: associations with differing cognitive operations
Biol. Pschyiatry
(2005) - et al.
Functional–anatomic correlates of subject priming in humans revealed by rapid presentation event-related fMRI
Neuron
(1998) - et al.
Functional analysis of genetic variation in COMT: effects on mRNA, protein and enzyme activity in postmortem brain
Am. J. Hum. Genet.
(2004) - et al.
The computational role of dopamine D1 receptors in working memory
Neural Netw.
(2002) - et al.
Frontal and temporal dysfunction of auditory stimulus processing in schizophrenia
NeuroImage
(2002) - et al.
Association of the G1947A COMT (Val(108/158)Met) gene polymorphism with prefrontal P300 during information processing
Biol. Psychiatry
(2003) Cellular basis of working memory
Neuron
(1995)- et al.
Improved optimisation for the robust and accurate linear registration and motion correction of brain images
NeuroImage
(2002)