To Each His Own Fear: Gender-Related Association of Anxiety, Substance Use, and Eating Disorders in a Representative Birth Cohort Sample of Young Adults with Either COMT Val158Met allele

An increase in the prevalence of mental disorders in young people is observed across several European countries [1]. According to the 2019 Global Burden of Disease study on mental health conditions in the 85 million young people aged 10–24 years living in 31 European countries, the most frequent mental illness was anxiety disorders [2]. While anxiety disorders are the most prevalent in both males and females, substantial differences in prevalence rates of most of psychiatric disorders occur in relation to gender: anxiety, depression, and eating disorders more often occur among females, while in males, alcohol, and drug use disorders are more common. These often mentioned gender-wise dissociations indirectly suggest that while all these psychiatric disorders have genetic background, the specific genetic mechanisms behind each disorder are strongly dependent on how environmental pressures impact on developmental processes in the brain. Successful reconciliation of findings in molecular genetics and basic neuroscience probably requires the consideration of adaptive mechanisms that respond to expression of “risk variants” of genes [3]. Along these lines, one hypothesis that concerns the very limited association with gene variants that have good theoretical background in affective neuroscience has posited these to serve as “plasticity genes,” with variants amplifying the effects of both negative and positive aspects of the environment [4, 5]. Any such genotype-dependent plasticity effect is bound to be dynamic, depending on the phases of development in the overall environmental context, and therefore, no allele would act as the “vulnerability allele” universally under all conditions [6]. Furthermore, the executive processes that respond to the environment, being connected by feedback loops with the motivational/affective circuits, would strive toward optimal adaptive response to any given environmental challenges, bringing agency, the motivated active engagement with the environment, into the equation [7]. In this view of the gene-environment-action model, the true role of key genes in major neurochemical systems may remain hidden without proper sample stratification.

A fair example of a genetic polymorphism relevant to brain function but unexpectedly without a clear role to play in psychiatry is served by the case of catechol-O-methyltransferase (COMT). COMT is an enzyme involved in the inactivation of dopamine, noradrenaline, and adrenaline, and this major role of COMT in catecholamine metabolism has led to the investigation of its variants in the genetic etiology of different psychiatric disorders, including psychotic, affective, and anxiety disorders [8]. Of the COMT gene polymorphisms, the single nucleotide polymorphism (guanine [G] vs. adenine [A]; rs4680) in codon 158 for the membrane-bound isoform of COMT, leading to methionine (Met) or valine (Val) alternation, has been most studied owing to its well-characterized functional implications at the biochemical level. Thus, the methionine-containing variant significantly reduces enzyme activity in the brain, leading to higher synaptic dopamine levels [9]. The polymorphic COMT enzyme activity has a trimodal distribution (high activity in Val/Val genotype, intermediate activity in Val/Met genotype, and low activity in Met/Met genotype), and overall, the difference in COMT activity can be three- to fourfold depending on this genotype (Val/Val vs. Met/Met) [10]. The genotype effect on dopamine levels is particularly evident in the prefrontal cortex, a brain region involved in emotional regulation and cognitive processing, where dopamine transporter levels are low and negative feedback via autoreceptors less effective [11]. In theory, these differences in dopamine metabolism should have implications for various psychiatric disorders that involve dysregulation of the dopamine system.

Indeed, associations between the COMT Val158Met polymorphism and the risk for multiple psychiatric conditions, including anxiety disorders [12, 13], eating disorders [14, 15], and substance use disorders [16], have been reported. Regarding anxiety disorders, the Met-allele of the COMT Val158Met polymorphism, resulting in lower enzyme activity and higher synaptic dopamine, has been associated with an increased risk for generalized anxiety disorder (GAD) [13], and in men with obsessive-compulsive disorder [17, 18]. On the other hand, Val-allele carriers may have a higher risk for panic disorder [19, 20], but this association has, in some studies, been found to hold only for Caucasians or for female subjects [8, 21, 22]. Studies conducted in American Indians and in Taiwan have shown that the COMT Val158Met polymorphism is also associated with an increased risk for alcohol and substance use disorders, with individuals carrying the Val-allele having a higher risk [23, 24]; a meta-analysis confirmed this for Asian samples [8]. In Caucasians, on the other hand, early-life adversity has been found to contribute to drinking and drug use, particularly in Met-allele carriers [16, 25]. Several studies have suggested the role of COMT Val158Met polymorphism in the etiology of eating disorders [14, 15, 26], but meta-analyses have failed to affirm these findings for anorexia nervosa [27] or bulimia nervosa and anorexia nervosa combined [28]. Taylor [8], however, in his meta-analytical review, reported a trend for anorexia nervosa, in which younger samples were associated with an overrepresentation of the Met-allele. While COMT was considered a candidate gene for schizophrenia, the meta-analytic approach has rejected the role of this genotype [29]. Neither have genome-wide association studies hit at COMT for any of these disorders [30‒32], but this is similar to other genes encoding proteins that are directly involved in neurotransmitter metabolism and action.

The existence of multiple but inconsistent psychiatric findings associating the COMT Val158Met genotype which has a clear functional significance may be explained, apart from the common notion of involvement of Type 1 errors, by the multiplicity of molecular pathways to a given disorder that vary by the type of gene-environment interactions involved in any specific sample. In the longitudinally studied ECPBHS sample, which is highly representative of the selected birth cohorts in a region, we have previously observed that the COMT Val158/Met polymorphism was associated with basic working memory and emotion perception functions [33, 34]. We had also observed that COMT Val/Val homozygosity was associated, in females, with a stepwise increase in neuroticism from childhood to early adulthood [35]. Therefore, we hypothesized that COMT Val158/Met variants can be found associated with specific psychiatric disorders in this population, especially with anxiety disorders, and that these associations at least partially depend on gender. Hence, we explored the relationship between COMT Val158Met and the incidence of psychiatric disorders in this highly representative birth cohort sample by 25 years of age.

The study sample was formed for the European Youth Heart Study (EYHS; 1998/1999) and later incorporated into the Estonian Children Personality Behaviour and Health Study (ECPBHS). It is an ethnically homogeneous sample of Caucasian subjects, and the rationale and procedure of sample formation and follow-up studies have previously been described [36, 37]. In brief, all schools of Tartu County, Estonia, which agreed to participate (54 of the total of 56), were included in the sampling using the probability proportional to the number of students of the respective age groups in the school, and 25 schools were selected. In 1998/99 (EYHS, later renamed ECPBHS owing to much expanded scope of the study), all children from grades 3 and 9 were invited to participate, and written informed consent was received from 79% of the requested subjects and their parents (original n = 1,238, including 569 males and 669 females). The present study included 936 participants from both study cohorts who participated in the psychiatric interview while at the age of 25 years (413 males, 523 females). Before participation, written informed consent was obtained from the participants. The study was approved by the Ethics Review Committee on Human Research of the University of Tartu and conducted in accordance with the Declaration of Helsinki.

The occurrence of psychiatric disorders by the age of 25 years was determined with the Mini-International Neuropsychiatric Interview (M.I.N.I.5.0.0) [38]; Estonian version [39] according to DSM-IV criteria by experienced clinical psychologists [40]. This interview was conducted with 936 participants (male n = 413; female n = 523).

The genotypes for the COMT Val158Met were determined from blood samples. Genotyping reactions of the COMT Val158Met polymorphism (rs4680) were performed in a total volume of 20 μL with 10–50 ng of template DNA as previously described [35]. The real-time polymerase chain reaction was performed with primers and fluorescent probes obtained from Applied Biosystems (Foster City, CA, USA) Custom TaqMan SNP Genotyping Assays. Real-time polymerase chain reaction components and final concentrations were as follows: 1:5 5 × HOT FIREPol Probe qPCR Mix Plus (ROX) (Solis BioDyne) and 1:20 80 × TaqMan Primers Probe (F 5′-CCC​AGC​GGA​TGG​TGG​AT-3′; R 5′-CAG​GCA​TGC​ACA​CCT​TGT​C-3′; Reporter 1-TTCGCTGGCATGAAG (VIC); Reporter 2-TCGCTGGCGTGAAG (FAM)). Reactions were performed on the ABI 7500 Real-Time PCR System, and the amplification procedure consisted of an initial denaturation step at 95°C for 15 min and forty 15 s cycles at 95°C and 1 min cycles at 60°C. All genotyping reactions were carried out in duplicates, and extra negative controls were added to each reaction plate. No inconsistencies occurred. All subjects were successfully genotyped. As expected, genotypes were found to be in the Hardy-Weinberg equilibrium.

Descriptive statistics was used for the frequency distribution of different psychiatric disorders in the sample. We used χ² tests to assess genotype distributions between psychiatric conditions. Two-tailed estimations of asymptotic significance were used. At first, the occurrence of disorders was examined by three genotypes, and then comparisons were made between homozygotes and heterozygotes (Val/Val with Met-allele carriers and Met/Met with Val-allele carriers of COMT Val158Met). The analyses were also carried out separately for males and females. The level of significance was set at p < 0.05. Binary logistic regression analysis with bootstrapping was used in order to explain the relationship between gender, COMT Val158Met genotype, and the occurrence of specific psychiatric disorders. Data were analyzed using IBM SPSS Statistics 28 (Chicago, IL, USA).

The lifetime occurrence of psychiatric disorders in 25-year-old males and females in the ECPBHS sample is presented in Table 1. Affective disorders and anxiety disorders were more common in females. Substance use disorders were significantly more prevalent among males and mostly present as alcohol use disorder.

Both Val- and Met-alleles of the COMT Val158Met polymorphism were associated with specific psychiatric disorders (Table 2). Across the whole sample, the Met-allele carriers had a significantly higher occurrence of agoraphobia (χ² = 4.10; p = 0.04) than Val/Val homozygotes.

Also, Met/Met homozygotes had a higher occurrence of agoraphobia (10 of 265) if compared to Val/Val subjects (1 of 187; χ² = 4.84; p = 0.03), while the comparison of Val/Val and Val/Met genotypes closely missed the conventional statistical difference (χ² = 3.43; p = 0.06). A binary logistic regression analysis was performed to ascertain the effects of gender and COMT Val158Met genotype on the likelihood of agoraphobia by the age of 25 years. The logistic regression model was statistically significant, χ² (df = 2, N = 936) = 15.3; p < 0.001. The model explained 7.4% (Nagelkerke R²) of the variance in occurrence of agoraphobia. Females (OR = 1.48, 95% CI [0.6, 3.2]) and Met-allele carriers of COMT Val158Met (OR = 1.87, 95% CI [0.6, 18.1]) were more likely to experience agoraphobia by the age of 25 years, p < 0.005 and p = 0.038, respectively.

The occurrence of panic disorder was significantly higher in female Met-allele carriers than in Val/Val homozygotes (χ² = 4.62; p = 0.03). In contrast, the occurrence of GAD was higher in Val/Val females when compared to the Met-allele carriers (χ² = 4.16; p = 0.04). We found no statistically significant association between the COMT Val158Met genotype and the occurrence of other anxiety-/stress-related disorders or mood disorders, except the higher prevalence of posttraumatic stress disorder in Val/Val males, but this was based on a single case. Given the above results, it was not surprising that no difference was found for COMT Val158Met genotypic distribution in the general categories “any anxiety disorder” (Table 2) or “any mood disorder” (data not shown), either in males or in females.

The number of subjects with psychotic disorders, as well as males with eating disorders, was too low for meaningful analysis. Female Val/Val homozygotes had a higher occurrence of eating disorders than Met-allele carriers of the COMT Val158Met polymorphism (χ² = 10.39; p = 0.001; Table 3).

COMT Val158Met Met/Met homozygotes had a significantly higher occurrence of alcohol use disorders (71/26.8% of 265) and any substance use disorder (75/28.3% of 265) compared to the Val-allele carriers (141/21% of 671 and 150/22.4% of 671; χ² = 3.62 and 3.68, respectively; p < 0.05). These trends were similar for males and females (data not shown). The regression model χ² (df = 2, N = 936) = 106.9, p < 0.001 revealed that male sex (OR = −1.63, 95% CI [−1.97, −1.33]) was a stronger predictor of any substance use disorder at the age of 25 years than Met/Met genotype of COMT Val158Met (OR = −0.3, 95% CI [−0.65, 0.05]), p < 0.001 and 0.08, respectively.

In a sample highly representative of two birth cohorts of a geographical region, we have revealed multiple associations of the COMT Val158Met polymorphism with psychiatric disorders. Associations were found for either allele. Irrespective of gender, carriers of the Met-allele had a higher prevalence of agoraphobia, while females made a larger contribution toward this association. Association of the genotype with panic disorder and GAD was present only in females and the genetic risk was similar to agoraphobia in the former case but opposite in the latter. Female Val/Val homozygotes had a higher prevalence of eating disorders. Met/Met homozygotes had a higher prevalence of substance (largely alcohol) use disorders, with a similar trend in males and females. Thus, Val/Val homozygosity increases the risk of GAD and eating disorder in females, while the Met-allele increases the risk of panic disorder and agoraphobia in females, and alcohol use disorder in either gender, the pattern of results in this sample bringing to mind the quote from The Jungle Book of Rudyard Kipling in the title of this manuscript.

The onset of common psychiatric conditions typically occurs during late adolescence and early adulthood [41], and thus, our assessment at the age of 25 years has probably captured most of the incidence of psychiatric disorders in this birth cohort representative sample. To our knowledge, no study has reported the association of COMT Val158Met genotype with multiple psychiatric disorders within the same sample. While attempting replication, it may be relevant to consider that the present sample is highly representative of regional birth cohorts, as it has been acknowledged that participation in (genetic) studies carries its own genetic bias [42]. Furthermore, dynamic gene × environment interactions make such results potentially sensitive to birth cohort effects [43, 44]; however, in the present study, the two birth cohorts appeared as subject to similar associations between COMT Val159Met alleles and psychiatric disorders. But even if the outcome of the present study were idiosyncratic to the setting where it was conducted, it would remain noteworthy that both COMT Val158Met alleles had specific psychiatric associations, and double dissociation by gene allele and gender was observed. This raises the question of what is different about (prefrontal) dopaminergic neurotransmission in males and females, and in panic disorder versus GAD or in eating disorders versus substance use disorders. Overall, the genetic correlation of eating disorders with anxiety disorders is very low and, if anything, negative with alcohol use disorder [45] and gender.

Dopaminergic systems have been implicated in many of the symptoms discussed in relation to anxiety disorders, including fear generalization [46]. The inactivating effect of COMT is particularly significant in the prefrontal cortex and amygdala [47], and the Met-allele of the COMT Val158Met polymorphism is associated with a significantly lower activity of the COMT enzyme, with correspondingly increased dopaminergic activity at synapses [48]. Thus, what is known of the functional significance of the COMT Val158Met polymorphism suggests that lower prefrontal dopaminergic signaling can predispose to general anxiety disorder and eating disorder, while a higher dopamine signal rather supports the incidence of agoraphobia, panic disorder, and alcohol use disorder. Structural and functional brain imaging studies have often implicated prefrontal cortex in GAD but not in panic disorder [49]; on the other hand, prefrontal cortex is involved in response of panic disorder to therapy [50]. The neurobiological emotion theory of Panksepp has posited that (general) anxiety and panic derive from evolutionarily distinct neurobiological circuits in the subcortical brain, with very different neurochemical underpinnings [51]. That broader networks beyond prefrontal cortex should be considered is suggested by the findings of COMT Val158Met-dependent differences in brain activation by unpleasant stimuli over extensive limbic areas and attentional network [52].

At the molecular level, male and female patients of psychiatric disorders can differ significantly, and this includes gene expression in prefrontal cortex. For example, Seney and colleagues [53] found a very limited number of overlaps between male- and female-depressed patients with regard to differentially expressed genes, and of these overlapping genes, differences with patients were in males and females mostly in opposite directions. COMT belongs to genes that could contribute to differences in brain function between males and females [54], and in human brains, its levels in the prefrontal cortex were shown to be higher in males than in females [9].

While altered dopaminergic neurotransmission is generally implicated in eating disorders, the exact nature of the alterations is not clear [55], and gender differences in the neurobiology of these disorders have, until recently, remained minimally studied. However, many differences in cerebral functional connectivity have been observed between males and females with eating disorders by use of a variety of seed regions [56].

Alcohol use disorder was associated with the COMT Met-allele (homozygosity) in this sample, which is compatible with some previous studies [16, 25]. Alcohol and other substance use disorders were fourfold more common among males, but genotype proportions with the disorder appeared similar in both genders. Prefrontal regions such as the dorsolateral prefrontal cortex and anterior cingulate cortex, affected by the COMT Val158Met variation, are implicated in excessive alcohol drinking, with more pronounced effects in young men [57]. In a prefrontal function-dependent task, alcohol-dependent patients performed worse, apomorphine-induced dopamine receptor stimulation reduced performance, especially in healthy controls, and COMT Met-allele carriers were more sensitive to the effect of apomorphine, while the homozygotes were the most sensitive [58].

Several limitations must be borne in mind: The sample size of this longitudinal study is limited, and thus, no correction for conducting multiple exploratory statistical comparisons was attempted; this is a limitation, but nevertheless, the number of sizeable differences between females and males, often nominally significant, is worth noting and requires further investigation in samples of similar nature. The diagnosis was made retrospectively for those subjects who did not meet diagnostic criteria during the evaluation but reported such a mental condition in the past: Our previous studies [40] have, however, provided support to the validity of the method. Another limitation is the lack of information in this type of a study about the exact onset time of mental disorders, as it would be of interest to examine the role of reaching sexual maturity; nevertheless, the size of the sample would likely not permit such an analysis.

In a regional highly birth cohort representative sample, either COMT rs4680 variant was observed in association with specific psychiatric disorders, and these associations were, in part, gender-dependent. Further research should identify the mediating mechanisms for the contribution to specific vulnerability or resilience by this plasticity gene.

The authors are grateful to the ECPBHS participants and the whole ECPBHS Team.

The study protocol was reviewed and approved by the Ethics Review Committee on Human Research of the University of Tartu, Estonia (license numbers: 49/30, 151/11, 197T-14, and 235/M-20) and was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all study participants.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

This work was supported by the Estonian Research Council (PRG1213).

Evelyn Kiive: conceptualization, data curation, formal analysis, investigation, methodology, and writing – review and editing. Triin Kurrikoff: investigation, methodology, and writing – review and editing. Toomas Veidebaum: conceptualization, funding acquisition, project administration, resources, supervision, and writing – review and editing. Jaanus Harro: conceptualization, funding acquisition, investigation, methodology, project administration, resources, supervision, and writing – review and editing.

The data that support the findings of this longitudinal study are not publicly available as they contain information that could compromise the privacy of research participants, but are available from the corresponding author [J.H.] who is the Chair of the Data Sharing Committee of ECPBHS [[email protected]] upon reasonable request.