Poster #16 - Evidence of an indirect effect of polygenic risk scores for childhood externalizing behavior through negative parenting behaviors

• In Event: 2-055 - Poster Session 06
  In Poster Session: PS 06 Section - Biological Processes: Neuroscience and Genetics

Integrative Statement

Background: Externalizing behaviors (EXT) include a constellation of aggressive/antisocial behaviors, such as callous and unemotional traits, oppositionality, and hyperactivity and inattention. Although genes play a key role in its etiology (Hicks et al., 2004), individual genetic variants only explain a small portion of its overall variance. Researchers are increasingly using polygenic approaches to optimally characterize the landscape of common genetic variation underlying EXT (Salvatore et. al., 2015). Still, very little is understood about how polygenic influences eventuate into EXT. In this study, we focused on parenting behaviors, including positive and negative dimensions, as possible mechanisms by which genes eventuate into EXT in children, especially given robust associations of parenting on childhood EXT along with the previously established passive gene-environment correlations (O'connor, Deater-Deckard, Fulker, Rutter, & Plomin, 1998). We hypothesize that parenting behaviors will mediate the association between polygenic risk and EXT.

Methods: Participants were 210 children (57.1% male, mean age=6.02, S.D.=.43) with and without EXT problems. A factor score for EXT was calculated from attention-deficit/hyperactivity disorder and oppositional defiant disorder symptoms from a fully structured clinical interview, and from the callousness dimension from the Inventory for Callous-Unemotional Traits. Factor scores for positive and negative parenting were calculated from positive and negative parenting items from the Alabama Parenting Questionnaire and from a survey about prosocial/positive parent-child dynamics (Eccles, personal communication). Polygenic Risk Scores (PRS) for EXT were calculated from a meta-analytic genome-wide association study of aggressive behavior in children (N=18,988; Pappa et. al., 2015). A GWAS p-value threshold of 0.96 explained the greatest proportion of variance in EXT in our sample and was used in the PRS. We conducted a multiple mediation model to test the indirect effects of PRS on EXT through the effects of positive and negative parenting, controlling for child’s race-ethnicity, sex, household income, and parental marital status.

Results: Figure 1 provides results of each regression path in the mediation model. Without accounting for the effects of parenting behavior, PRS was marginally associated with EXT (b=.0535, s.e.=.0291, p=.0679). After accounting for the mediators, PRS was no longer significantly associated with EXT (b=.0343, s.e.=.0284, p=.2287). The 95% confidence interval (CI) for the specific indirect effect of PRS on EXT through the effects of negative parenting behavior did not include zero, however the 95% CI for positive parenting behavior did (95% CIs =.0039, .0939 and -.0713, .0044, respectively).

Conclusions: Results provide new evidence (i.e., via PRS) to support the notion that genetic influences for EXT are correlated with the environment (i.e., negative parenting), which in turn, is correlated with childhood EXT (i.e., a passive or possible evocative gene-environment correlation effect a la Knafo and Jaffee, 2013). Previous studies that have examined PRS-environment interactions for EXT involving parenting behaviors may be confounded by this correlation. Future analyses will examine longitudinal associations between parenting behavior and childhood EXT, providing temporal separation between the mediator and outcome to establish causality of these effects.

Authors

• Bailey Ann Immel, University of Wisconsin- Madison
  Presenting Author

• Alexis Hosch, University of Wisconsin- Madison
  Non-Presenting Author

• James J. Li, University of Wisconsin-Madison
  Non-Presenting Author