- Top of page
- Participants and methods
The brain-derived neurotrophic factor (BDNF) Val66Met allelic variation is linked to both the occurrence of mood disorders and antidepressant response. These findings are not universally observed, and the mechanism by which this variation results in increased risk for mood disorders is unclear. One possible explanation is an epistatic relationship with other neurotransmitter genes associated with depression risk, such as the serotonin-transporter-linked promotor region (5-HTTLPR). Further, it is unclear how the coexistence of the BDNF Met and 5-HTTLPR S variants affects the function of the affective and cognitive control systems. To address this question, we conducted a functional magnetic resonance imaging (fMRI) study in 38 older adults (20 healthy and 18 remitted from major depressive disorder). Subjects performed an emotional oddball task during the fMRI scan and provided blood samples for genotyping. Our analyses examined the relationship between genotypes and brain activation to sad distractors and attentional targets. We found that 5-HTTLPR S allele carriers exhibited stronger activation in the amygdala in response to sad distractors, whereas BDNF Met carriers exhibited increased activation to sad stimuli but decreased activation to attentional targets in the dorsolateral prefrontal and dorsomedial prefrontal cortices. In addition, subjects with both the S allele and Met allele genes exhibited increased activation to sad stimuli in the subgenual cingulate and posterior cingulate. Our results indicate that the Met allele alone or in combination with 5-HTTLPR S allele may increase reactivity to sad stimuli, which might represent a neural mechanism underlying increased depression vulnerability.
Depression has been hypothesized to be a failure of executive control over emotion (Disner et al. 2011; Mayberg 1997), a theory largely supported by neuroimaging studies, which find decreased function of the executive system and increased activity of the emotional system (Siegle et al. 2007; Steele et al. 2007). Genetic variants may contribute to the imbalanced executive and affective systems resulting in increased vulnerability to developing major depressive disorder (MDD). Among the best-studied genes in MDD are variants of the serotonin-transporter-linked promoter region (5-HTTLPR) and brain-derived neurotrophic factor (BDNF) genes (Caspi et al. 2003; Gatt et al. 2009). However, it is unclear how BDNF and 5-HTTLPR variants independently and interactively influence neural systems associated with affective and executive function.
The Met variant of the BDNF Val66Met polymorphism results in decreased activity-dependent secretion of BDNF (Chen et al. 2004; Egan et al. 2003) and is associated with deficits in hippocampal function (Chen et al. 2006). Although some studies found that the Met allele is more common in geriatric depression (Taylor et al. 2007), meta-analyses concluded that Val allele was linked to anxiety and unrelated to depression (Frustaci et al. 2008; Verhagen et al. 2010). The BDNF gene also exhibits epistatic interactions with other genes (Gatt et al. 2009; Levinson 2006), including variants in 5-HTTLPR, which could partly explain the inconsistent findings in Met BDNF variant. The short (S) allele variant of 5-HTTLPR is associated with reduced transcription of 5-HTT (Lesch et al. 1996). Compared with long (L) allele homozygous individuals, healthy S allele carriers exhibit an exaggerated amygdala response to fearful and angry faces (Hariri et al. 2002, 2005). Supporting the theory of epistatic effect between BDNF and 5-HTTLPR genes, some studies reported that the combination of the S and Met alleles predicts depressive symptoms (Kaufman et al. 2006; Kim et al. 2007; Wichers et al. 2008). Others, however, suggested that the Met allele may protect against the influence of the 5-HTTLPR S allele on neural systems (Pezawas et al. 2008). Overall, little is known about the interactive effects of these genes on neural function.
Neuroimaging studies examining the BDNF Val66Met gene have predominately focused on memory-related hippocampal function (Egan et al. 2003; Hariri et al. 2003; Kanellopoulos et al. 2011; Pezawas et al. 2004) with little attention on emotional processing (Montag et al. 2008). Conversely, studies examining 5-HTTLPR neglected the executive function and only focused on emotion. Moreover, studies on emotion used mixed negative stimuli or fearful/angry stimuli (Hariri et al. 2005; Heinz et al. 2007), making it difficult to tease apart the genetic effects on depression from anxiety. Investigating neural responses to exclusively sad stimuli should be more sensitive in revealing the functional contribution of the genes specific to depressed mood.
In this study, we tested for both main and epistatic effects of the BDNF and 5-HTTLPR polymorphisms on neural activation related to sad emotion and executive function in a group of older adults. We hypothesized that epistatic interactions of the BDNF Met and 5-HTTLPR S genes exaggerate affective and executive dysfunction.
- Top of page
- Participants and methods
We found that the 5-HTTLPR S allele carriers had exaggerated amygdala activation to sad stimuli without alteration in neural responses during target detection. Met allele carriers showed increased activation when processing sad information and decreased activation during target detection in the dlPFC and dmPFC, regions typically associated with cognitive processing (Wang et al. 2005). Furthermore, those carrying both the S allele of the 5-HTTLPR gene and Met allele of the BDNF gene had stronger activation to sad stimuli in the PCC and the subgenual cingulate than those who did not carry both genetic variants. On the other hand, there was no significant group difference in the S & Val/Val allele carriers compared with non-(S & Val/Val) carriers. Taken together, the results indicate that the Met allele of the BDNF gene rather than the Val allele is associated with affective and executive dysfunction, which might increase vulnerability to depression.
Our finding that the 5-HTTLPR S allele carriers had hyperactivity to sad stimuli in the amygdala, a key structure of the ‘emotional system’, is consistent with majority of studies in the literature (Bertolino et al. 2005; Hariri et al. 2002, 2005; Munafo et al. 2008). Some studies argued that the strong amgydala activation in the S allele carriers was because of reduced activation to neutral stimuli rather than increased activation to negative stimuli (Canli et al. 2005a; Heinz et al. 2007). Heinz and colleagues (2007) found increased amygdala activation to fixation vs. neutral contrast in the S allele carriers and they emphasized that a passively resting baseline might increase anxiety as an aversive stimuli to the S allele carriers. In this study, the baseline was phase-scrambled pictures. Relative to this baseline, we did not find decreased activation to neutral stimuli in the S allele carriers (Fig. 1, the bar graph), and the increased amygdala activation was specific to sad stimuli. Previous studies used threatening pictures (Bertolino et al. 2005), mixed emotional pictures (Hariri et al. 2002) or negative words (Canli et al. 2005b) as emotional stimuli. Our study extended these previous findings to sad emotion in older subjects with S allele polymorphism.
Literature examining the association between 5-HTTLPR genetic variants and neural function was primarily focused on responses to negative stimuli. The predominant findings of studies are increased amygdala activation and amygdala-vmPFC (ventromedial prefrontal cortex) connectivity (Pezawas et al. 2005), along with increased perfusion in the amygdala and vmPFC during resting state in one study (Rao et al. 2007). Studies of S allele carriers in MDD also found increased amygdala activation when perceiving masked (Dannlowski et al. 2008) or non-masked emotional pictures (Dannlowski et al. 2007; Gillihan et al. 2011). Studies examining whether the 5-HTTLPR genetic variants affect the executive system are rare. In this study, we investigated the modulation effect of 5-HTTLPR genetic variants on neural responses during both emotional and executive processing. We found that the S/S, S/L and L/L genotypes were not different in response to attentional targets. These results suggest that S-allele-associated genetic vulnerability predominantly influences the affective system rather than the executive system.
Consistent with studies reporting decreased hippocampal volume and function, we found decreased activation of targets in the Met allele carriers in the right hippocampus/parahippocampus region. The studies on emotional processing in BDNF Met allele carriers reported increased activation in the amygdala during an affective startle reflex paradigm (Montag et al. 2008); increased activation in the anterior cingulate, brainstem and insula during fear processing (Mukherjee et al. 2011); and decreased activation in the orbitofrontal cortex, striatum and amygdala to negative stimuli (Gasic et al. 2009). Similarly, we found an increased response to sad distractors but a decreased response to targets in the dlPFC and dmPFC in Met allele carriers relative to the Val homozygous carriers. The dlPFC typically shows deactivation during a ‘passive viewing’ task (Wang et al. 2005, 2008b; Yamasaki et al. 2002). Activation, instead of deactivation, to sad stimuli in the dlPFC and dmPFC in the Met allele carriers suggests increased cognitive processing rather than passively processing sad emotional information. The dmPFC has been related to self-referential processing or an implicit cognitive aspect of emotional processing such as judgment of emotion (Northoff & Bermpohl 2004) or attentional modulation of emotion (Bermpohl et al. 2006; Northoff et al. 2006). Therefore, the Met allele carriers might have initiated a cognitive coping strategy to actively deal with sad distractors, which might have been more distracting to them compared to the Val/Val carriers, while passively viewing sad stimuli. The altered activation to both sad and target stimuli suggest that the Met allele can alter neural function in both emotional and executive systems, which predisposes individuals to MDD.
While 5-HTTLPR S alleles increased activation to sad stimuli in the amygdala, a combination of the BDNF Met with 5-HTTLPR S alleles increased activation in broader regions including the sgACC and PCC, regions that are critically associated with depression (Drevets 2001; Mayberg 2009). The result is not surprising because BDNF Met may exacerbate monoamine deficiencies at the genetic level (Ren-Patterson et al. 2005) and enhance hyperactivity in the emotional system because of executive dysfunction at the neural level when 5-HTTLPR S allele is present. This is consistent with the genome-wide association (GWA) study, which revealed that individuals carrying 5-HTTLPR S/S and the BDNF Met variants scored high on neuroticism (Terracciano et al. 2010). It is also consistent with the findings that individuals carrying the S/S and BDNF Met genes had the highest depression scores (Kaufman et al. 2006; Kim et al. 2007; Wichers et al. 2008) and high rumination in life stress (Clasen et al. 2011). Therefore, our results support the hypothesis that the combination of the BDNF Met with 5-HTTLPR S genes increases one's risk for depression, which opposes the alternative explanation that it has a protective effect on depression (Martinowich & Lu 2008; Pezawas et al. 2008).
One counterintuitive finding in our study was the trend of greater degree of residual symptoms (high MADRS score) as the increase in numbers of 5-HTTLPR L alleles, although importantly this trend did not achieve statistical significance. Notably, all our subjects were in a fully remitted state; thus, our results do not reflect the degree of antidepressant response but could be related to residual symptoms. One potential explanation for this trend is that remitted L/L allele carriers may be at increased risk for residual depressive symptoms than are S/S allele carriers. However, to our knowledge there is no evidence from the literature supporting this theory. Carefully designed studies are needed to further address this issue.
The significant limitation of this study is the sample size, which does not allow us to separate our study sample into remitted and control groups in order to further investigate an interaction or mediation effect between the BDNF Val66Met and 5-HTTLPR genetic variants. Given that the significant differences between the remitted and control groups in our previous study (i.e. decreased activation in the remitted group in the dACC and aPCC to attentional targets (Wang et al. 2008a) were not overlapping the findings of the genetic variants reported here, it is less likely that our results were dominantly affected by depression history. Another limitation of this study is mixed medication in some remitted subjects. We did not find significant medication difference neither (Table 1) among the S/S, S/L and L/L genotype groups nor between the BDNF Met/Val and Val/Val carrier groups, which can partially cancel out the medication effect on brain activation. Although not always found, generally speaking antidepressants reduce activation in the amygdala and increase activation in the dlPFC (Robertson et al. 2007; Siegle et al. 2007). Therefore, one might expect stronger results if all subjects were not medicated with antidepressants. Nevertheless, future studies are needed in large samples free of medication.
It is also noteworthy that we did not find any difference in task performance among genetic variants during conducting the emotional oddball task. Also, we could not prove that the changes in the emotional and executive systems increased susceptibility to depression from our study. The increased susceptibility to depression that we discussed above was only based on changes in neural activation, and those with risk genes had changed activation in the same direction as those with depression. Future studies using more mental challenging tasks to evoke executive dysfunction and longitudinally following up on risk-gene carriers are needed to confirm whether activation changes in the emotional and executive systems confers increased vulnerability to depression.
In summary, results from our current study suggest that the Met allele rather than Val allele is possibly a risk gene for depression. Furthermore, our results also indicate that a combination of the S allele of the 5-HTTLPR gene and Met allele of the BDNF genotypes may have increased depression vulnerability supported by exaggerated activation in the PCC and the subgenual cingulate in response to sad stimuli. Future studies comparing sad and fearful or anger stimuli might further elucidate the genetic impact on neural function related to depression vs. anxiety disorders.