- Top of page
- Materials and methods
Studies in children have shown that the genetic influence on cognition is positively correlated with socioeconomic status. Catechol-O-methyltransferase (COMT) Val158Met, a common, functional polymorphism, has been implicated in executive cognition and working memory. Imaging studies have shown that the variant Met allele is associated with more efficient prefrontal cortical processing and better attention but also emotional vulnerability to stress. We hypothesized that COMT Val158Met genotype would interact with years of education (yrs ed), one indicator of socioeconomic adversity, to predict cognitive task performance. We therefore administered the Wechsler Adult Intelligence Scale-Revised (WAIS-R) to 328 community-derived, genotyped, Plains American Indians (mean yrs ed = 12; range = 5–18). We found significant genotypic effects on WAIS-R measures of long-term memory, working memory and attention. The Met allele was associated with improved performance in the Information and Picture Completion subscales; Met/Met homozygotes performed the best. COMT genotype interacted with yrs ed to influence Information and Block Design scores: Met allele carriers’ scores improved markedly with increasing yrs ed, whereas the scores of Val/Val individuals were only marginally influenced by yrs ed. There was a crossover of effects at 11–12 yrs ed: in the less educated group, Met allele carriers actually performed worse than Val/Val individuals perhaps because of emotional vulnerability to educational adversity, but in the better educated group, Met allele carriers excelled. Our study in Plains American Indians has shown that COMT Val158Met influences several aspects of cognition and some of its effects are moderated by educational adversity.
Convergent evidence from animal studies and human neuroimaging studies indicate that dopamine (DA) plays a major role in executive processing and enhances the cortical physiological signal to noise ratio (reviewed in Tunbridge et al. 2006). Catechol-O-methyltransferase (COMT) is largely responsible for the metabolism of DA and norepinephrine in the prefrontal cortex (PFC). Studies in COMT-deficient mice have shown that half of the DA decline in the PFC results from COMT-mediated enzymatic degradation (Yavich et al. 2007). A functional polymorphism, COMT Val158Met, responsible for a fourfold difference in enzyme activity in humans, (Chen et al. 2004; Weinshilboum et al. 1999) is present at high frequencies in all ethnic groups (Palmatier et al. 1999).
COMT Val158Met has so far only been found in humans and it is therefore of interest that this polymorphism has been implicated in cognition, a behavioral domain in which humans differ from nonhuman primates. COMT Val158Met modulates PFC function; individuals with the ancestral high-activity Val allele perform less well on tests of working memory and executive cognition than individuals with the low-activity Met allele, and Val allele load predicts a less efficient physiological response in the PFC (Barnett et al. 2007; Egan et al. 2001; Hariri & Weinberger, 2003; Tunbridge et al. 2006). COMT inhibitors have been shown to improve working memory and attention in animals (Tunbridge et al. 2004) and also in humans (Apud et al. 2007). The Met allele has been associated with a more anxious, cautious personality (Enoch et al. 2003; Hettema et al. 2008; Olsson et al. 2005) and with greater activation in response to negative emotional stimuli in the limbic system and connected prefrontal areas (Smolka et al. 2005).
Several studies in young children and adolescents have shown that the genetic influence on cognition is positively correlated with socioeconomic status; that is, the shared environment accounts for more of the variance in cognitive skills in disadvantaged children, whereas the genetic contribution is greater in children from affluent families, and the influence of heritability on childhood cognitive skills increases with parental education (Harden et al. 2007; Rowe et al. 1999; Turkheimer et al. 2003). We were interested to see whether COMT Val158Met genotype would interact with years of education (yrs ed), one possible indicator of socioeconomic status, to predict cognitive skills in adults. We conducted a study using subscales from the Wechsler Adult Intelligence Scale-Revised (WAIS-R) (Wechsler 1981) that are indicators of various aspects of cognition including working memory, long-term memory, numerical reasoning, attention and concept formation. The participants in our study were Plains American Indians from rural Oklahoma with a high prevalence of lifetime Diagnostic and Statistical Manual of Mental Disorders, third edition, revised (DSM-III-R) alcoholism who came from a deprived socioeconomic environment that included limited educational opportunities. We hypothesized that the COMT Val158Met polymorphism was likely to have an effect on cognitive skills assessed by the WAIS-R and, based on the previously described studies, that the effect might be moderated by yrs ed. Moreover, the effects might be modified by alcoholism.
- Top of page
- Materials and methods
Our study sample consisted of 328 Plains American Indian men and women, 71% of whom had not completed high school and were therefore considered educationally deprived. We found significant main effects of COMT and interactive effects of COMT and yrs ed on cognition accessed by three of the seven WAIS-R subscales: Inf, PC and BD. COMT Val158Met genotype had a significant effect on Inf and PC. The Met allele was associated with higher scores for both tests in a dose-dependent fashion. Secondary analyses revealed that this significant COMT main effect derived from the more educated group (≥13 yrs ed) for Inf but from the less educated group (< 13 yrs ed) for PC. From Table 2, it can be seen that scores for Inf, an indicator of verbal comprehension and long-term memory, improved for all genotypes with increasing education, but more markedly for Met allele carriers. In contrast, Val/Met individuals in both the high- and low-education groups had essentially the same scores for PC, an indicator of attention to fine detail and perceptual organization, whereas Val/Val and Met/Met individuals benefited from increased yrs ed. Our results indicate that the effects of COMT Val158Met on cognition may differ, depending on the domains that are being tested.
Figure 1a shows that Inf scores for Met/Met and Val/Met individuals improved markedly with increasing education, whereas Val/Val individuals were less influenced by yrs ed. Also, Fig. 1 shows a crossover effect at approximately 11 yrs ed: Met/Met and Val/Met individuals with less than 11 yrs ed actually performed worse than Val/Val individuals but the reverse was the case in individuals with more than 11 yrs ed. There was a similar interaction effect on BD (working memory) scores with a crossover of effect at approximately 12.5 yrs ed (Fig. 1b). There were also apparent interactive COMT × yrs ed effects on the Sim (concept formation) and Dsp (working memory) tests (Fig. 1c,d), but these were more modest (P = 0.047–0.011) than the effects on Inf and BD (P = 0.007–0.005) and may have been because of rGE (see below). For Sim and Dsp, the Val/Met heterozygotes were more sensitive to the influence of education than either homozygote. The Val/Met heterozygotes’ scores increased with yrs ed at the same rate for all four tests. Although the Val/Val homozygotes’ scores also increased with yrs ed at the same rate for all four tests, this rate was much lower than that of the heterozygotes (Fig. 1c,d). The crossover point between Val/Val and Val/Met individuals in all four tests was at 11–12 yrs ed. In contrast, the rate of increase for the Met/Met homozygotes differed between tests: it was in the expected direction for Inf, BD and Sim scores but for Dsp, Met/Met homozygotes unexpectedly derived no benefit from increased yrs ed. One explanation for this result may be the small sample size of Met/Met individuals (n = 26). Indeed, the significance of all G×E interactions improved considerably when Met/Met and Val/Met individuals were counted as one group. Another possibility may be the influence of rGE, discussed below. Finally, it is also plausible that not all tasks are sensitive to the same variables.
We found evidence of rGE. Met allele dosage was significantly associated with increased yrs ed although this only accounted for 4% of the variance. Possible explanations for this rGE might be that if Met carriers are cognitively more gifted they may be more motivated to stay on in education. Perhaps the more cautious, anxious, inhibited personality of a Met carrier may also contribute to perseverance in education. It is known that G×E interactions are difficult to assess in the presence of rGE (Rutter & Silberg 2002). When we repeated the analyses after excluding the 33 individuals with 5–9 yrs ed who had the lowest proportion of Met alleles, rGE was no longer significant; however, the G×E interactions remained significant for the Inf and BD tests but not for the Sim and Dsp tests. This suggests that these are true G×E interactions influencing Inf and BD scores, but the G×E results for Sim and Dsp might have been influenced by rGE.
As discussed in the Introduction, a large body of literature has shown that COMT activity is an important modulator of dorsolateral PFC-dependent cognitive tasks (Apud et al. 2007; Egan et al. 2001; Hariri & Weinberger 2003; Malhotra et al. 2002; Tunbridge et al. 2006). We found that one (BD) of the three WAIS-R tests that access working memory was influenced by COMT genotype interacting with yrs ed, a marker for educational adversity. Moreover, another working memory test, DSym, showed a trend main effect for COMT in the predicted direction: the Met allele was associated with higher test scores. COMT Val158Met has also been shown to modulate attention through the cingulate cortex, located in the PFC. The Val allele has been linked with poorer attentional control coupled with greater activity in the cingulate cortex and the effects of Val allele load increase as attentional demands increase (Blasi et al. 2005). Attentional demands may be greater in less educated individuals and our finding that COMT genotype had the strongest impact on PC scores in the less educated group appears to fit in here.
A network of brain regions is known to be involved in the encoding and retrieval of memory including two linked regions, the hippocampus and the ventrolateral PFC (VLPFC). DA is an important modulator of their interactions (Schacter & Wagner 1999). COMT is also widely expressed in the hippocampus. Carriers of the Val158 allele perform less well on tests of long-term memory (de Frias et al. 2004). During encoding and retrieval of a recognition memory task, the COMT Val158 allele has been associated with poorer performance, decreased hippocampal activation together with less efficient VLPFC activation and impaired functional coupling between these two regions (Bertolino et al. 2006). Our results for the WAIS-R Inf test are supported by these earlier studies.
There are few published studies that have looked at the influence of COMT genotype on WAIS subscales. In a study of psychotic individuals, the COMT Val158 allele was associated with greater deterioration over time in Inf, DSym and Vocabulary scores (Mata et al. 2006). In healthy, elderly individuals, the Val/Val genotype has been linked with diminished cognition, particularly for the DSym test (for which we found a trend main effect) (Starr et al. 2007) and the Met/Met genotype has been linked with better performance on a measure of attention (Liu et al. 2008). Our results agree with these earlier findings.
The COMT Met158 allele has been associated with greater activation to negative emotional stimuli in the limbic system, hippocampus and VLPFC (Drabant et al. 2006; Smolka et al. 2005). Met/Met homozygotes show increased frontal coupling of the limbic and prefrontal regions (Drabant et al. 2006). The Met allele association with heightened reactivity and connectivity in corticolimbic circuits may reflect a predisposition for inflexible processing of affective stimuli and a tendency for rumination that may account for the Met allele association with increased anxiety and emotional dysregulation noted in earlier studies (Drabant et al. 2006; Hettema et al. 2008; Olsson et al. 2005), including in our sample (Enoch et al. 2003). Thus, the crossover effect that we have noted in Fig. 1 may indicate a balance between emotional resilience/poorer cognitive skills (Val allele) and stress-vulnerability/superior cognition (Met allele). One possible interpretation of our G×E interaction results for the Inf and BD subtests is that, in good educational environments, Met allele carriers (Met/Met and/or Val/Met individuals) may be able to develop their full potential for cognitive skills that are superior to those of Val/Val homozygotes. However, when Met allele carriers are exposed to significant educational adversity, their emotional vulnerability may overwhelm their ability to develop cognitive skills. In contrast, Val/Val homozygotes may be more resilient to educational adversity.
Our sample had a high prevalence of lifetime AUD (73% in men and 47% in women) and therefore AUD was included as a between-subject factor. AUD had a modest effect on PC scores in the total sample (P = 0.02) and on BD scores in the less educated group (P = 0.02). However, there was no interactive effect with COMT genotype.
There are a few limitations to our study. We did not correct for multiple testing. We analyzed seven WAIS-R subscales with no a priori hypothesis as to which test might have the strongest association; therefore, a Bonferroni corrected P value would be P = 0.007. Several, but not all, of our association results achieved that degree of statistical significance. Moreover, when analyses were conducted with Met/Met and Val/Met individuals included as one group to allow for the small Met/Met sample size, the significance level improved considerably. Nevertheless, the results from our study should be considered preliminary until reproduced in other data sets. The effects of the COMT Val158Met polymorphism on the WAIS-R tests were not huge: COMT genotype accounted at best for 5% of the variance in test scores in the Plains Indians. This is similar to the variance (4%) previously found for COMT influence on perseverative errors in executive cognition tests (Egan et al. 2001). Numerous other factors also influence cognition, such as the quality of education, parental, socioeconomic and cultural factors. Other genes, for example, brain-derived neurotrophic factor, may also influence cognitive function. In this study of American Indians, we used the WAIS-R subtests that are considered to be culturally appropriate. Nevertheless, our results should be interpreted with caution until they have been replicated in other ethnic groups with similar educational adversity.
In conclusion, our study in Plains American Indians has shown that COMT genotype has an effect on measures of cognition that access several brain regions including the dorsolateral and ventrolateral PFC, the cingulate cortex and the hippocampus. These effects were detected within the broad range of cognitive tests that form the WAIS-R. Met allele carriers generally performed better on tests of cognition, particularly when raised in better learning environments, but Val/Val individuals appeared to be more resilient when exposed to educational adversity.