Report

COMT val158met Genotype Affects µ-Opioid Neurotransmitter Responses to a Pain Stressor

See allHide authors and affiliations

Science  21 Feb 2003:
Vol. 299, Issue 5610, pp. 1240-1243
DOI: 10.1126/science.1078546

Abstract

Responses to pain and other stressors are regulated by interactions between multiple brain areas and neurochemical systems. We examined the influence of a common functional genetic polymorphism affecting the metabolism of catecholamines on the modulation of responses to sustained pain in humans. Individuals homozygous for themet158 allele of the catechol-O-methyltransferase (COMT) polymorphism (val158met) showed diminished regional μ-opioid system responses to pain compared with heterozygotes. These effects were accompanied by higher sensory and affective ratings of pain and a more negative internal affective state. Opposite effects were observed in val158 homozygotes. The COMTval158met polymorphism thus influences the human experience of pain and may underlie interindividual differences in the adaptation and responses to pain and other stressful stimuli.

Strong interest exists for the discovery of genes that cause individual differences in responses to physical and environmental challenges. In the case of pain, both sensitivity and inhibition are traits that vary considerably among individuals, with some of the variability being attributed to genetic factors (1, 2). However, the influence of genes on regulatory processes in the human brain is particularly difficult to resolve. A functional genetic variant may affect not only the protein coded by the gene in question but may also have downstream effects contributing to the overall system response. Furthermore, differences in human resiliency and stress responses determine individual vulnerabilities to many psychiatric and other complex diseases (3, 4).

Here, we focus on an abundant functional polymorphism of the cathechol-O-methyltransferase (COMT) gene that codes the substitution of valine (val) by methionine (met) at codon 158 (val158met). This substitution is associated with a difference in thermostability leading to a three- to fourfold reduction in the activity of the COMT enzyme (5). The alleles are codominant so that individuals with theval/val genotype have the highest activity of COMT, those with the met/met genotype have the lowest activity of COMT, and heterozygous individuals are intermediate. Theval158met genotypes have been linked to a number of behavioral diseases of complex etiology (6–10).

COMT is one of the enzymes that metabolizes catecholamines, thereby acting as a key modulator of dopaminergic and adrenergic/noradrenergic neurotransmission (11, 12). Different levels of COMT activity conferred by val158met genotypes may then have important influences on functions regulated by these neurotransmitters, including μ-opioid system responses. In animal models, the chronic activation of dopaminergic neurotransmission and D2 receptors, a situation parallel to that encountered inmet/met homozygotes, reduces the neuronal content of enkephalin peptides and induces compensatory increases in regional μ-opioid receptor concentrations in various brain regions. Reductions in D2 receptor–mediated neurotransmission, similar to that achieved by the higher levels of COMT activity in val/val homozygotes, results in opposite effects on the μ-opioid system (13–16). Therefore, we hypothesized that chronic overactivity of the dopaminergic system induced by the low-function met/met COMT enzyme would be associated with a lesser capacity to activate μ-opioid neurotransmission under provocation conditions by virtue of a lower neuronal content of enkephalin. Compensatory increases in μ-opioid receptor binding should also be observed under these circumstances. We hypothesized intermediate effects in heterozygous individuals, while the presence of the higher metabolic capacity of the val/val COMT genotype would be associated with higher enkephalin content, a superior capacity to activate μ-opioid neurotransmission and possibly compensatory reductions in receptor binding levels. The μ-opioid system is activated in response to stressors, pain, and other salient environmental stimuli, typically reducing pain and stress responses (1, 17–22). At a psychophysical level, we expected the variations in the magnitude of μ-opioid system activation induced by the COMT genotypes to result in varying capacities to suppress responses to pain and other stressors.

To elucidate the contribution of val158met COMT genotypes on the functional responses of the μ-opioid system and their psychophysical correlates, we exposed human subjects to a sustained pain challenge. As even moderate levels of pain become temporally more sustained, it becomes a significant physical and emotional stressor (23) and activates suppressive μ-opioid neurotransmitter responses (1, 24). This activation is then associated with reductions in sensory and affective qualities of pain and in the negative affective state of the volunteers (1, 24).

We studied 29 healthy volunteers (15 men and 14 women, 20 to 30 years of age) genotyped with respect toval158met polymorphisms (24) with positron emission tomography (PET) and the μ-opioid receptor-selective radiotracer [11C]carfentanil (24, 25). We studied women during the early follicular phase of their menstrual cycle, when estradiol and progesterone levels are lowest (estradiol, 38.2 ± 25.0 pg/ml; progesterone, <0.7 ng/ml in all cases) (25). This minimizes possible effects of estradiol on COMT gene transcription (26). We measured receptor binding in vivo (binding potential, BP) (25) twice: during intensity-controlled sustained pain induced by the infusion of small amounts of 5% hypertonic saline into the masseter muscle and during the infusion of nonpainful 0.9% isotonic saline. Subjects were blinded and randomized as to the order of the conditions. Pain intensity was maintained constant and between 35 and 45 visual-analog-scale (VAS) intensity units with a computer-controlled infusion system and feedback from subjects on their present pain intensity (25). Pain-related sensory and affective qualities were measured with the McGill pain questionnaire (MPQ) (27). We assessed the internal emotional state with the positive and negative affectivity scale (PANAS) (28). We calculated the magnitude of activation of μ-opioid receptor–mediated neurotransmission by subtracting BP values during pain from baseline (nonpainful control) BP. Under the experimental conditions used, the activation of μ-opioid neurotransmission in vivo is detected as reductions in the BP measure from control to pain conditions (1,24, 25).

From the total sample of 29 volunteers, we selected 18 as matched on the basis of gender and scan order to reduce the effects of these variables (11 met/val, 4 met/met, 3val/val). The 11 heterozygotes matched themet/met sample and 9 of them also matched theval/val group.

We tested the main effect of the val158metgenotype on μ-opioid system activation and μ-opioid receptor BP maps by one-way analysis of variance (ANOVA) applied on a voxel-by-voxel basis (25).

We detected significant effects of genotype on μ-opioid system activation (degrees of freedom = 2, 15 for all regions,P < 0.05 after correction for multiple comparisons) in the anterior thalamus [x, y, zcoordinates (millimeters), 5, −1, −2; F = 29.3], the thalamic pulvinar ipsilateral to the painful challenge (x,y, z, −8, −24, 8; F = 17.0), and the ventral basal ganglia bilaterally (peak x,y, z coordinates, 9, −1, −12; F= 27.7) (Fig. 1). The latter encompassed the nucleus accumbens and adjacent ventral pallidum and extended caudally into the subthalamic nucleus. We observed possible effects of genotype, at significance levels approaching the multiple comparisons threshold, in the ipsilateral dorsal thalamic nuclei (x, y,z, −14, −19, 18; F = 8.3), the amygdala ipsilateral to pain (x, y, z, −25, −5, −22; F = 8.9), and the amygdala and anterior temporal cortex contralateral to pain (x, y,z, 28, −6, −21; F = 8.9 and 27, −6, −10;F = 11.6, respectively) (P values < 0.0001, uncorrected).

Figure 1

Effect of COMT val158metgenotypes and associated COMT activity on μ-opioid system BP and μ-opioid receptor system activation in response to sustained pain. (Upper left) Three-dimensional display of significant ANOVA results for the effects of COMT val158metgenotypes on baseline μ-opioid receptor BP. Significant effects of genotype on baseline BP were observed in the anterior and posterior (pulvinar ipsilateral to pain) thalamus (1). Near the multiple comparisons threshold, possible effects were observed bilaterally in the nucleus accumbens and ventral pallidum and in the contralateral thalamic pulvinar. (Lower left) ANOVA results for the effects of COMT val158met genotypes on μ-opioid system activation during sustained pain stress. Significant effects were observed in the anterior and posterior (pulvinar) thalamus (1) and striatopallidal regions [nucleus accumbens (2), ventral pallidum (3), subthalamic nucleus, bilaterally]. ANOVAF scores are represented by the pseudocolor scale between the two images. (Upper right) Correlations between COMT activity conferred by the COMT val158metpolymorphisms and baseline μ-opioid receptor BP. COMT activity was coded as follows: –1, met/met; 0, val/met; +1, val/val. Significant correlations were obtained in the anterior thalamus (1) (x, y, z, 0, −12, 4; z = 4.56) and the nucleus accumbens (2) and adjacent ventral pallidum (3) ipsilateral to pain (x,y, z, 12, 5, −9; z = 4.33). Possible relationships approaching the multiple comparisons threshold were observed in the contralateral nucleus accumbens/ventral pallidum (x, y,z, −5, 5, −8; z = 4.11) (P< 0.0001 uncorrected) . Mean ± SD of nucleus accumbens (ipsilateral) BP values are shown (upper right inset). (Lower right) Correlations between COMT activity and μ-opioid system activation were obtained in the posterior thalamus (1) ipsilateral to the challenge (−7, −21, 4; z = 4.74), nucleus accumbens (2), and ventral pallidum (3), bilaterally (5, 5, 1, and −7, 5, 5; z = 5.00 and 6.10), and the contralateral amygdala and adjacent temporal cortex (peak coordinates 27, −6, −11; z = 4.78). The subthalamic nuclei, bilaterally, registered levels of statistical significance approaching the multiple comparisons threshold (10, −4, −4 and −11, −1, −5; z = 4.22 and 4.02;P < 0.0001 uncorrected). Mean ± SD values for the percent change in the magnitude of μ-opioid system activation in the nucleus accumbens are shown (lower right inset).Z scores for the correlations are represented by the pseudocolor scale between the two right images. Threshold for display of results was set at P < 0.01 for all analyses.

We observed significant effects of genotype on baseline μ-opioid receptor BP values in the anterior thalamus (x,y, z, −1, 11, 3; F = 18.9) and ipsilateral thalamic pulvinar (x, y,z, −14, −24, 6; F = 12.3) (Fig. 1). We noted possible genotype effects on BP approaching multiple comparison thresholds in the ventral basal ganglia, bilaterally (nucleus accumbens, ventral pallidum) (x, y, z, 11, 7, −10; F = 11.9) and the contralateral thalamic pulvinar (x, y, z, 17, −26, 2;F = 8.8) (P < 0.0001 uncorrected).

Because the effects of genotype were hypothesized to occur in particular directions (that is, COMT activity correlating positively with μ-opioid system activation and negatively with μ-opioid receptor binding), three levels of COMT activity were then introduced as a main covariate of interest in the SPM analysis. We observed significant correlations in the predicted directions for regions that overlapped with those identified by the ANOVAs. These included the anterior thalamus, nucleus accumbens, and ventral pallidum for μ-opioid binding and the thalamic pulvinar, nucleus accumbens, ventral pallidum, and amygdala for μ-opioid system activation (Fig. 1, fig. S1).

Post hoc two-sample, two-tailed t-tests were then applied for contrasts between the homozygous samples and their individually matched heterozygous comparison groups. These analyses confirmed that the magnitude of μ-opioid system activation in response to the pain challenge was lower in low COMT function met/met subjects (n = 4) than in the matched, intermediate COMT functionmet/val sample (n = 11). Significant differences were obtained in striatopallidal pathway nuclei and the interconnected amygdala (Fig. 2, table S1). Effect sizes for these group differences in μ-opioid system activation ranged from 15% to 32%. Areas approaching the multiple comparisons threshold included the rostral anterior cingulate (x, y, z, 3, 27, −8;z = 3.68), anterior insula (x, y,z, 34, 16, −9; z = 3.48), and thalamic pulvinar (x, y, z, −16, −29, 1;z = 3.48) (P < 0.0001 uncorrected).

Figure 2

Differences in μ-opioid receptor binding in vivo and μ-opioid neurotransmitter system activation in response to sustained pain between met/met and matchedval/met subjects. (Upper row, left) Three- dimensional representation of μ-opioid receptor BP values in a representative healthy volunteer, superimposed over a magnetic resonance image standardized to ICBM stereotactic coordinates. BP values are represented by the pseudocolor scale in the lower part of the figure. Coronal views. Anterior to posterior (y axis) ICBM coordinates are shown in red. (Upper row, right) Significantly higher baseline μ-opioid receptor BP in the ventral pallidum (y = 1), extending anteriorly and medially into the nucleus accumbens inmet/met individuals (n = 4), compared with met/val subjects (n = 11). (Lower row) Brain areas where lower magnitudes of μ-opioid system activation were detected inmet/met individuals (n = 4) compared with matched met/val subjects (n = 11). From anterior to posterior (right to left side of the figure), coronal cuts show the following: y = 7, posterior area of nucleus accumbens and ventral pallidum, extending into hypothalamus;y = 0, ventral pallidum, subthalamus, amygdala;y = −2, ventral pallidum, subthalamus, amygdala.Z scores of statistical significance are represented by the pseudocolor scale in the lower part of the image and are superimposed over an anatomically standardized magnetic resonance image in coronal views. Left side of the figure corresponds to the side contralateral to the pain challenge, and right side corresponds to the side ipsilateral to the pain challenge. The threshold for display of results was set atP < 0.01 for all analyses.

We also confirmed increases in μ-opioid receptor BP inmet/met subjects compared with the met/val group. These were observed in the ventral basal ganglia unilaterally, centered in the ventral pallidum, and extending medially and anteriorly to the nucleus accumbens (Fig. 2) (x, y, z, 18, 1, −7; z = 4.52). BP increases in themet/met group ranged from 32% to 41% compared with heterozygotes (mean ± SD, met/met 2.0 ± 0.6;met/val 1.4 ± 0.4). These effects were regionally specific and did not reflect global changes in μ-opioid receptor binding, because whole brain BP values did not differ between groups (met/met 0.58 ± 0.05; met/val 0.58 ± 0.08).

Identical post hoc comparisons were conducted betweenval/val individuals (n = 3) and their matched met/val heterozygous comparison group (n = 9). We confirmed significantly higher regional μ-opioid system activation in the high COMT functionval/val group compared with the matched met/valsample. The regions involved included the dorsal anterior cingulate, anterior thalamus, and cerebellar vermis (Fig. 3, table S1). Effect sizes for these group differences in μ-opioid system activation ranged from 16% to 30% for these regions. Approaching the multiple comparisons threshold, we noted val/val genotype effects in the same direction in the dorsomedial area of the thalamus ipsilateral to pain (x,y, z, −8, −20, 15; z = 4.14) and the periaqueductal gray (x, y, z, −1, −33, −9; z = 4.04) (P < 0.0001 uncorrected).

Figure 3

Differences in μ-opioid receptor binding in vivo and μ-opioid neurotransmitter system activation in response to sustained pain between val/val and met/valsubjects. Coronal views. Anterior to posterior (yaxis) ICBM coordinates are shown in red. (Upper row) Lower baseline μ-opioid receptor BP in the anterior thalamus (y = −11) and pulvinar area of the thalamus (y = −24) in val/val individuals (n = 3) compared with matched met/valsubjects (n = 9). (Lower row) Brain areas where higher magnitudes of μ-opioid system activation were detected in val/val individuals compared with met/valsubjects. From anterior to posterior brain areas, dorsal anterior cingulate and anterior thalamus (y = −2) and cerebellar vermis (y = −49).Z scores of statistical significance are represented by the pseudocolor scale in the lower part of the image and are superimposed over an anatomically standardized magnetic resonance image in coronal views. Left side of the figure represents the side contralateral to the pain challenge, and right side represents the side ipsilateral to the pain challenge. The threshold for display of results was set atP < 0.01 for all analyses.

We also observed significantly lower baseline μ-opioid receptor BP in the val/val group in the anterior thalamus (x,y, z, −1, −11, 2; z = 5.70) and thalamic pulvinar (x, y, z, −14, −24, 4; z = 5.05) (Fig. 3). Reductions approaching multiple comparison thresholds were also observed in the nucleus accumbens, bilaterally (x, y, z, 10, 9, −9, and −12, 10, −7; z = 3.63 and 3.42), and amygdala (x, y, z, −26, −1, −22; z= 3.81) (P < 0.0001 uncorrected). These corresponded to average reductions in BP in the val/valgroup of 26% in the anterior thalamus (range 15% to 40%), 33% in the pulvinar (range 19% to 41%), 33% in the nucleus accumbens (range 21% to 38%), and 26% in the amygdala (range 7% to 48%). Global, whole brain BP values again did not differ between groups (val/val 0.49 ± 0.12; met/val 0.53 ± 0.06).

These data confirm the hypothesis that varying levels of catecholamine metabolism induced by the COMTmet158val polymorphism are associated with downstream alterations in the functional responses of the μ-opioid neurotransmitter system and compensatory changes in μ-opioid receptor binding. We observed statistically significant effects in striatopallidal circuits receiving prominent dopaminergic input, confirming a significant role of COMT in the metabolization of basal ganglia dopamine in humans (29). The data also agree with observations in animal models, where the chronic activation or blockade of D2 receptor–mediated dopamine neurotransmission induced opposite changes in enkephalin and μ-opioid receptor concentrations (B max) in striatopallidal projections connecting the nucleus accumbens with the ventral pallidum (13,15, 16). Ventral pallidal μ-opioid receptors are thought to be centrally involved in the regulation of a motivational circuit that includes the nucleus accumbens, thalamus, amygdala, prefrontal cortex, and brainstem nuclei, which integrates sensory information and motor responses with affective and cognitive influences (30–34).

The COMT met158val polymorphism also influenced μ-opioid responses and receptor concentrations in several other regions, such as the dorsal anterior cingulate, anterior and posterior thalamus, amygdala, and cerebellar vermis. Dorsal anterior cingulate and anterior thalamic μ-opioid neurotransmission are implicated in the modulation of affective components of the pain experience (1, 35), whereas the posterior thalamus (pulvinar region) and the amygdala form part of a network involved in processing affective stimuli (36). The μ-opioid receptors in the dorsal anterior cingulate have been further involved in cognitive effects on pain processing, such as those that mediate placebo responses (20, 37). The amygdala has also been implicated in μ-opioid receptor–dependent, stress-, and fear-induced analgesia and emotional regulation (21,22, 38). We noted trends in the same directions in the rostral anterior cingulate, insular cortex, dorsal thalamus, and periaqueductal gray, although they were below multiple comparison thresholds.

As noted above, dopamine μ-opioid system interactions have been described in striatopallidal circuits, as well as in the rostral anterior cingulate and the amygdala (14,16). However, very low levels of dopaminergic innervation are present in the thalamus, whereas noradrenergic input is prominent in this region (39). μ-Opioid receptor regulation of noradrenaline release has also been described in the dorsal anterior cingulate and amygdala (38, 40). Therefore, it is likely that some of the COMTval158met effects on regional μ-opioid transmission are due to interactions, direct or indirect, with noradrenergic terminals.

These data demonstrate robust effects of the functional COMT val158met polymorphism on human μ-opioid receptor binding in vivo and on the capacity to activate μ-opioid neurotransmitter system responses. We then examined the contribution of the COMT val158met polymorphism to interindividual variations in the psychophysiological responses to the pain challenge in the entire sample of 29 subjects genotyped (table S2).

Pain intensity ratings for the initial 150-μl bolus of hypertonic saline and the average VAS pain intensity ratings for the duration of the study were similar across genotype groups. However, the volume of hypertonic saline necessary to reach and maintain the preset level of pain intensity (24), a measure of pain sensitivity, paralleled the COMT activity of the genotypes. It was lowest inmet/met, intermediate in met/val, and highest inval/val. This correlation achieved statistical significance for the volume of hypertonic saline required late (10 to 20 min) but not early (0 to 10 min) in the challenge (table S2). This is an important detail, in agreement with previous data demonstrating that the μ-opioid system becomes activated with prolonged but not acute stressors (17). Conversely, MPQ sensory, affective, and total scores and PANAS negative affect scores were highest in individuals with the lowest (met/met) COMT activity, followed by heterozygotes and val/val homozygotes. Statistically significant thresholds were reached for the negative correlation between COMT activity and the ratings of the negative internal affective state induced by the painful challenge (table S2).

This report shows that a common, functional genetic polymorphism of the human COMT enzyme causes unique regional neurochemical system activations that are paralleled by distinct psychophysical response traits during sustained pain. These data emphasize the need for a systems-level approach to the elucidation of neurobiological processes whereby genetic variation, neuronal functional measures (that is, endophenotypes), and phenotypic traits are fully integrated.

Supporting Online Material

www.sciencemag.org/cgi/content/full/299/5610/1240/DC1

Materials and Methods

Fig. S1

Tables S1 and S2

  • * To whom correspondence should be addressed. E-mail: zubieta{at}umich.edu.

REFERENCES AND NOTES

View Abstract

Navigate This Article