Abstract:

There has been a dramatic rise in gene-environment interaction (G × E) studies of depression over the last two decades. These studies are pivotal to understanding the etiology of depression and individual differences in environmental sensitivity. However, these studies rarely take into consideration how the genotype by environment interactions change over development and how the interactions work on the developmental trajectories of depression. The brain-derived neurotrophic factor (BDNF) gene is a good candidate for the investigation of the dynamic genetic effects on depression because it is involved in several age-related changes in behavior and brain maturation. On the one hand, the effect of the BDNF gene may depend on the basal BDNF level. The BDNF level peaks during adolescence, so the effect of the BDNF gene on depression may change during development. Moreover, peer experiences change may alter epigenetic modifications of the BDNF gene, which may change the pattern of gene-environment interactions. On the other hand, according to the developmental cascades model, the differences in genetic effects on depression may increase over time in that initial depressive symptoms may evoke poor peer experiences. Taken together, this study aimed to investigate the age differences in the G × E interaction on depression and the G × E effect on the developmental trajectories of depression.

One thousand and eighty-six adolescents (aged 11-12 years with a mean of 12.32, 50% girls) were followed up for three years. Saliva samples, self-reported depressive symptoms, and peer nomination were all collected. All of the measures showed good reliability. Concurrent hierarchical regression analyses and latent growth curve models (LGCMs) were conducted. We also completed re-parameterized regression and parallel LGCMs to understand the gene by environment interaction pattern and the dynamic association between peer relationships and depression.

Both peer rejection and peer acceptance showed significant correlations with depressive symptoms, concurrently and prospectively (as shown in Table 1). As shown in Table 2 and Table 3, the BDNF Val66Met polymorphism significantly moderated the influence of peer rejection—but not peer acceptance—on youth depressive symptoms at three time points; however, the susceptible genotype changed over time. In particular, the effect of peer rejection on depression was stronger in MetMet compared to ValMet carriers at 12 years of age; the effect of peer rejection on depression was stronger in MetMet and ValVal compared to ValMet carriers at 13 years of age; the effect of peer rejection on depression was stronger in ValVal carriers compared to ValMet carriers at 14 years of age (see Figure 1). The re-parameterized regression analyses showed that the pattern of the BDNF gene by peer rejection interaction also changed over time (Table 4, Table 5, and Table 6). The LGCMs suggested that adolescents’ depression increased in a linear trajectory from 12 to 14 years of age (Table 7). In addition, there were significant genotype differences in the change of depression over time, but this effect was not moderated by peer relationships (Table 7, Figure 2-4).

These findings may move research in the field away from the simplistic notion of risk alleles, recognizing that an allele may be a risk factor during one period and a protective factor during another. Further, this study has progressed the conceptualization of how genes and the environment interact to influence the developmental trajectories of depression during early adolescence.

Key words: adolescent depression, BDNF, peer rejection, age differences