Catechol-O-methyltransferase activity does not influence emotional processing in men

Introduction

Dopaminergic abnormalities are prominent in many psychiatric disorders and contribute substantially to their cognitive and emotional components (Figee et al., 2016; Goldman-Rakic et al., 2004). The catechol-O-methyltransferase (COMT) enzyme metabolises dopamine in the prefrontal cortex (PFC) (Byers et al., 2020; Tunbridge, 2004). The human COMT gene contains a functional polymorphism (Val¹⁵⁸Met; rs4680) that influences enzyme activity: the ancestral Val¹⁵⁸ allele has ~40% greater activity than the Met¹⁵⁸ allele (Tunbridge et al., 2019). COMT activity can also be altered pharmacologically by COMT inhibitors, licenced for the adjunctive treatment of Parkinson’s disease (Zürcher et al., 1990).

An inverted-U-shaped relationship exists between dopamine and PFC-dependent cognitive function, whereby either too little or too much dopamine signalling impairs cognitive performance (Goldman-Rakic, 2000; Williams and Goldman-Rakic, 1995). Separately, both COMT inhibition (with the brain-penetrant COMT inhibitor tolcapone) and the Met¹⁵⁸ allele are associated with improved PFC dopamine-dependent cognitive task performance, compared with placebo and the Val¹⁵⁸ allele, respectively (Barkus et al., 2016; Byers et al., 2020; Farrell et al., 2012; Gasparini et al., 1997; Tan et al., 2007; Troudet et al., 2016; Tunbridge, 2004). Furthermore, and consistent with the inverted-U relationship, COMT genotype and tolcapone have interactive effects: COMT inhibition improves cognition in individuals with high COMT activity (e.g. Val¹⁵⁸ homozygotes) but has less effect, or may even impair, cognition in those with low COMT activity (e.g. Met¹⁵⁸ homozygotes) (Barkus et al., 2016; Farrell et al., 2012; Giakoumaki et al., 2008). Given these findings, COMT inhibitors with more optimal properties are under active development for the treatment of cognitive dysfunction in schizophrenia and other disorders (Byers et al., 2020; Su et al., 2021). However, COMT’s significance for clinically relevant domains other than executive cognition remains largely unexplored (DeBrosse et al., 2020; Korn et al., 2021; Tunbridge et al., 2012) and may have significance for understanding potential additional indications or side effects for these novel compounds.

Several lines of evidence suggest that COMT may influence emotional processing. Dopamine modulates brain regions implicated in the processing of emotional sensory information including the PFC, basolateral amygdala and nucleus accumbens (Gordon-Fennell and Stuber, 2021; Grace and Rosenkranz, 2002; Karreman and Moghaddam, 1996), and dopamine agonism is a potential therapeutic approach for the treatment of mood disorders (Dunlop and Nemeroff, 2007; Romeo et al., 2018). Evidence from both human and animal studies demonstrates that COMT genotype influences aspects of emotional processing. Neuroimaging studies demonstrate consistent associations between the COMT Val¹⁵⁸Met polymorphism and brain activation (determined using fMRI) during emotional processing: the Met¹⁵⁸ allele is associated with greater limbic brain activation (including hippocampus, amygdala and PFC) in response to negatively valanced stimuli (Drabant et al., 2006; Mier et al., 2010; Rasch et al., 2010; Smolka, 2005; Smolka et al., 2007; Williams et al., 2010; although see Kempton et al., 2009, who found greater limbic activation in Val¹⁵⁸ vs Met¹⁵⁸ homozygotes). Caution should be exercised when extrapolating simple differences in fMRI signal to underlying psychological processes; nevertheless, these findings are broadly consistent with a model in which the Met¹⁵⁸ allele confers greater attention and/or sensitivity to negative emotional information, compared to the Val¹⁵⁸ allele. This hypothesis is consistent with reported associations between the Met¹⁵⁸ allele and risk for various anxiety and stress-related disorders (Domschke et al., 2004; Kolassa et al., 2010; Pooley et al., 2007). However, findings from behavioural studies are more mixed. Thus, while in some mouse models, lower COMT activity is associated with increased anxiety-like behaviours (Gogos et al., 1998; Papaleo et al., 2008), this is not seen for all COMT models (Barkus et al., 2016). Findings in human volunteers are also inconsistent: while some studies report exaggerated responses to negative stimuli and greater negative biases in processing emotional stimuli associated with the Met¹⁵⁸, compared with the Val¹⁵⁸, allele (Gohier et al., 2014; Gong et al., 2013; Montag et al., 2008), others have reported the opposite pattern (Weiss et al., 2007) or have found no genotype differences in emotional processing (Carrà et al., 2017; Defrancesco et al., 2011; Kempton et al., 2009).

To our knowledge, no studies have examined the impact of COMT inhibition on emotional processing in human participants or whether its effects interact with genotype (and, if so, whether this also follows an inverted-U). However, in our earlier study, we found that tolcapone increased visual analogue score ratings of ‘happiness’ in Val¹⁵⁸ but not Met¹⁵⁸ individuals (Farrell et al., 2012). Further hints of a potential impact of COMT inhibition on emotional processing come from a small, open clinical trial that provided preliminary evidence that tolcapone might have antidepressant potential (Fava et al., 1999).

Taken together, the available data suggest that COMT genotype influences limbic activation during emotional processing; however, effects on emotional behaviour are not clear-cut. Furthermore, although there are hints that COMT inhibition may positively impact on mood, these findings are preliminary and in need of replication. It is also not known if or how these effects differ based on COMT genotype. This study, therefore, aimed to comprehensively assess the impact of COMT genotype and COMT inhibition on emotional processing. To achieve this, we investigated the separate and interactive effects of COMT inhibition and Val¹⁵⁸Met genotype on emotional processing in healthy men, using an emotional test battery which tests multiple dimensions of emotional behaviours and is sensitive to antidepressant actions (Harmer et al., 2011).

Methods

We performed a double-blind, randomised, placebo-controlled, experimental medicine study that investigated non-smoking healthy males homozygous for the COMT Val¹⁵⁸Met polymorphism. The study’s design paralleled that of our study of tolcapone’s effects on working memory (Farrell et al., 2012).

Results

The demographics of the final sample are shown in Table 1. There were no robust effects of Val¹⁵⁸Met genotype, inhibition or their interaction on any of the emotional processing measures (Table 2). Thus, while there were main effects of emotion on accuracy, misclassifications and reaction time (Fs > 21; ps < 4.7 × 10^–37) during facial emotion recognition, there were no main or interactive effects of genotype or drug (Fs < 1.22; ps > 0.31). There was a main effect of valence on reaction time (F_1,69 = 32.0; p = 3.2 × 10^–7) but not accuracy (F_1,69 = 0.9; p = 0.35) during assessment of word valence but no other main or interactive effects (Fs < 0.61; ps > 0.05); similar findings were obtained for their later recall (main effect of valence for correct recalls: F_1,69 = 12.9; p = 0.001; main effect of valence for errors: F_1,69 = 6.0; p = 0.017; other effects: Fs < 0.55; ps > 0.47). Finally, other than a main effect of emotion (F_1,68 = 11.6; p = 0.001), there were no other main or interactive effects on attentional vigilance towards happy or fearful faces (Fs < 1.61; ps > 0.12).

OPEN IN VIEWER

Table 1.

Demographics of participants.

	Met Placebo	Val Placebo	Met Tolcapone	Val Tolcapone
Number	19	19	19	17
Age (years)	23.3 ± 1.0	25.3 ± 1.3	24.8 ± 1.0	22.5 ± 1.2
NART predicted Full Scale IQ a	117.1 ± 1.2	115.6 ± 1.8	115.9 ± 1.5	113.9 ± 1.7
Baseline State Anxiety b	28.2 ± 1.2	28.1 ± 1.2	31.4 ± 1.8	30.2 ± 1.1
Baseline Trait Anxiety b	30.6 ± 1.7	32.1 ± 1.4	34.4 ± 2.1	32.1 ± 1.4

NART: National Adult Reading Test.

Values are mean ± SEM.

a

Non-native speakers excluded.

b

Spielberger’s State-Trait Anxiety Inventory.

OPEN IN VIEWER

Table 2.

Behavioural measures of emotional processing measured using the Emotional Test Battery (ETB).

	Val Placebo	Met Placebo	Val Tolcapone	Met Tolcapone
Facial Expression Recognition Task (FERT)
Accuracy (%)
Anger	46.3 ± 2.3	50.0 ± 2.1	51.9 ± 3.0	46.7 ± 3.2
Disgust	42.5 ± 3.3	47.9 ± 3.0	44.2 ± 4.1	47.0 ± 4.0
Fear	43.3 ± 3.4	44.6 ± 4.1	45.2 ± 3.5	48.8 ± 3.6
Happy	68.5 ± 2.6	66.9 ± 2.0	63.1 ± 1.8	65.0 ± 3.1
Sad	44.6 ± 2.6	49.6 ± 2.3	50.0 ± 4.0	50.8 ± 3.6
Surprise	57.6 ± 2.0	55.7 ± 2.8	58.0 ± 2.1	54.5 ± 2.7
Neutral	68.9 ± 4.5	69.4 ± 4.3	71.9 ± 3.9	71.1 ± 3.8
Misclassifications (%)
Anger	14.5 ± 2.2	14.1 ± 1.7	13.1 ± 1.6	14.2 ± 1.2
Disgust	10.4 ± 2.0	7.3 ± 1.5	7.5 ± 1.5	9.0 ± 1.4
Fear	8.0 ± 1.3	8.2 ± 1.4	7.1 ± 1.1	7.1 ± 0.9
Happy	2.2 ± 0.6	3.4 ± 0.9	2.8 ± 0.6	5.5 ± 2.1
Sad	8.79 ± 1.4	9.94 ± 1.9	8.31 ± 2.2	8.89 ± 1.7
Surprise	10.31 ± 1.4	12.83 ± 1.8	11.00 ± 2.0	11.94 ± 1.6
Neutral	63.26 ± 4.1	67.33 ± 4.7	67.75 ± 4.6	64.89 ± 3.3
Reaction time (ms)
Anger	1798 ± 91	1820 ± 107	1795 ± 81	1776 ± 84
Disgust	1794 ± 120	1897 ± 113	1800 ± 67	2009 ± 113
Fear	2122 ± 75	2167 ± 117	2104 ± 144	2104 ± 107
Happy	1653 ± 64	1603 ± 53	1598 ± 51	1604 ± 74
Sad	1858 ± 71	1973 ± 98	1907 ± 87	1866 ± 91
Surprise	1907 ± 122	1822 ± 95	1817 ± 72	1951 ± 98
Neutral	1748 ± 126	1725 ± 149	1655 ± 116	1520 ± 92
Emotional Categorisation Task (ECAT)
Reaction time (ms)
Positive	1047 ± 64	1104 ± 50	1132 ± 42	1073 ± 51
Negative	1158 ± 59	1122 ± 37	1224 ± 54	1163 ± 58
Faces Dot Probe Task (FDOT)
Vigilance bias score (ms)
Unmasked positive	11 ± 9	−18 ± 8	−4 ± 9	−11 ± 7
Unmasked negative	6 ± 9	−4 ± 9	−1 ± 7	20 ± 11
Masked positive	−10 ± 10	−4 ± 11	−15 ± 7	−24 ± 10
Masked negative	10 ± 11	6 ± 10	9 ± 9	10 ± 7
Emotional Recall Task (EREC)
Hits (n correct)
Positive	4.0 ± 0.5	4.9 ± 0.6	4.0 ± 0.6	4.8 ± 0.5
Negative	3.1 ± 0.5	3.9 ± 0.4	3.5 ± 0.5	3.6 ± 0.5
Intrusions (n wrong)
Positive	2.2 ± 0.5	1.8 ± 0.4	1.7 ± 0.3	2.1 ± 0.4
Negative	1.7 ± 0.6	0.9 ± 0.3	1.3 ± 0.3	1.2 ± 0.6

Values are mean ± SEM.

Comparing pre- and post-testing timepoints, there were no main or interactive effects on change scores for subjective mood VAS ratings (i.e. happiness, sadness or anxiety: Fs < 0.43; ps > 0.51) or total mood disturbance (POMS) (Fs < 0.39; ps > 0.54; Figure 1). There was a main effect of drug on VAS drowsiness (F_1,69 = 4.2; p = 0.044; change in drowsiness greater in tolcapone than placebo) in the absence of other main or interactive effects (Fs < 3.0; ps > 0.089; Figure 1). However, there were no effects on VAS alertness ratings (Fs < 1.4; ps > 0.24; Figure 1). Finally, there was a drug × genotype interaction for VAS nausea ratings (F_1,69 = 4.4; p = 0.039) in the absence of other main effects (Fs < 1.1; ps > 0.31), due to a genotype difference (Val > Met; p = 0.03) in the placebo group that was not observed in the tolcapone group (p = 0.44).

OPEN IN VIEWER

Figure 1.

Changes in subjective ratings in the different genotype and drug groups. The change scores from pre–post drug timepoints are shown broken down by group. Boxes indicate the median, 25th and 75th percentile values, with the whiskers indicating the maximum and minimum values. Only self-rated drowsiness and nausea differed between groups (see text for details).

Discussion

We found no effect of either COMT inhibition or COMT Val¹⁵⁸Met genotype on performance on an emotional test battery or any effects on subjective mood ratings. Thus, our findings suggest that, at least in healthy men, COMT does not influence behavioural measures of emotional processing. To our knowledge, ours is the first study to examine the effect on COMT genotype on multiple aspects of emotional processing and mood within a single cohort of human volunteers as well as the first to assess the impact of acute COMT inhibition on these processes. Given the similarities in experimental design and study population, the current data suggest that our previous finding of a positive effect of tolcapone on subjective happiness ratings represents a false positive (Farrell et al., 2012), consistent with a lack of effect of COMT genotype on subjective mood ratings reported elsewhere (Gibbs et al., 2014).

Detection of facial emotions is the area of emotional behaviour that is best studied with respect to COMT genotype, although a variety of different experimental approaches have been employed. Our findings agree with most earlier studies showing no overall effect of COMT genotype on facial emotion detection (Carrà et al., 2017; Defrancesco et al., 2011; Kempton et al., 2009; Soeiro-de-Souza et al., 2012; Tamm et al., 2016). We cannot rule out more subtle effects of COMT on detection of specific types of emotional information. However, it is notable that where more subtle differences have been found, results are inconsistent. For example, while some studies found genotype effects that were specific to negative emotions (Lischke et al., 2019; Weiss et al., 2007), other studies report altered detection of neutral, but not emotional, facial expressions associated with COMT (Erkoreka et al., 2020; Gohier et al., 2014).

Our negative behavioural findings are of interest in the context of the better-studied relationship between COMT genotype and brain activation during emotional processing. Thus, although there are some conflicting reports (Bishop et al., 2006; Domschke et al., 2012), most studies (including by meta-analysis) suggest that the Met allele is associated with greater limbic activation during emotional processing (Drabant et al., 2006; Mier et al., 2010). There is also some evidence that COMT genotype influences patterns of functional connectivity during emotional processing (Opmeer et al., 2013; Vai et al., 2017; but see Surguladze et al., 2012). Notably, this disconnect between neural measures and behavioural effects has been observed within subjects, both in the specific case of COMT genotype (Kempton et al., 2009) and after other manipulations of dopamine function (Martens et al., 2021). It is possible that that measurements of quantitative biological traits, such as BOLD response, are more sensitive to the subtle effects of genetic variation than studies looking at overt behaviour (Meyer-Lindenberg and Weinberger, 2006). Alternatively, in the case of genetically encoded differences such as COMT Val¹⁵⁸Met that presumably affect enzyme activity throughout development (Tunbridge et al., 2007), it is possible that differences observed at the neural level represent compensatory changes in the precise circuitry used to achieve a given behavioural outcome. It may be that such subtle differences in ‘wiring’ are insufficient to lead to any observable behavioural differences in healthy controls under baseline conditions. It is therefore notable that there is some evidence, albeit inconsistent, for associations between COMT genotype and emotion-related behaviours that are modulated by environmental adversity (Asselmann et al., 2018; Evans et al., 2009; Lovallo et al., 2017). Given the difficulty in disentangling such effects in epidemiological studies, it will be of interest to investigate the impact of stressors and other challenges on the relationships between COMT activity and emotional processing under controlled laboratory conditions.

Taken together, our data suggest that variation in COMT activity, whether genetic or pharmacological in origin, has no demonstrable effect on mood and emotional processing in healthy men under baseline conditions. Nevertheless, given the design and demographics of this study, further research is required to investigate the extent to which our findings generalise to other populations, to establish how COMT-related differences in neural processing relate to emotional behaviours and to assess whether prolonged COMT inhibition affects emotional processing. Our findings do not have immediate therapeutic relevance, since the licenced COMT inhibitors are either not brain-penetrant or are limited by hepatotoxicity. However, our results suggest that the novel, brain-penetrant COMT inhibitors under development for cognitive dysfunction (Byers et al., 2020) are unlikely to have major off-target effects on mood and related behaviours, at least when administered acutely.