Association Between Genetic Risks for Obesity and Working Memory in Children

Abstract

Introduction: Obesity is highly heritable, and recent evidence demonstrates that obesity is associated with cognitive deficits, specifically working memory. However, the relationship between genetic risks for obesity and working memory is not clear. In addition, whether the effect of these genetic risks on working memory in children is mediated by increased body mass index (BMI) has not been elucidated. Methods: In order to test whether the polygenic risk score (PRS) for obesity in adulthood (adulthood-BMI-PRS) is associated with working memory at 8 years of age, and whether the effect is mediated by childhood BMI, in children from the general population, participants in the Hamamatsu Birth Cohort for Mothers and Children (HBC) study in Hamamatsu, Japan, underwent testing for association of adulthood-BMI-PRS with working memory. HBC data collection began in December 2007 and is ongoing. Adulthood-BMI-PRS values were generated using summary data from the recent genome-wide association study (GWAS) undertaken in Japan, and the significance of thresholds was calculated for each outcome. Outcomes measured included the working memory index (WMI) of Weschler Intelligence Scale-4 (WISC-IV) scores and the BMI at 8 years of age. Gene-set enrichment analysis was conducted to clarify the molecular basis common to adulthood-BMI and childhood-WMI. Mediation analysis was performed to assess whether childhood-BMI of children mediated the association between adulthood-BMI-PRS and working memory. Results: A total of 734 participants (377 males, 357 females) were analyzed. Adulthood-BMI-PRS was associated with lower childhood-WMI (β[SE], -1.807 [0.668]; p = 0.010, corrected) of WISC-IV. Gene-set enrichment analyses found that regulation of neurotrophin Trk receptor signaling (β[SE], -2.020 [6.39]; p = 0.002, corrected), negative regulation of GTPase activity (β[SE], 2.001 [0.630]; p = 0.002, corrected), and regulation of gene expression epigenetic (β[SE], -2.119 [0.664]; p = 0.002, corrected) were enriched in BMI in adulthood and WMI in childhood. Mediation analysis showed that there is no mediation effect of childhood-BMI between the adulthood-BMI-PRS and working memory deficits in children. Conclusion: Adulthood-BMI-PRS was associated with working memory among children in the general population. These genetic risks were not mediated by the childhood-BMI itself and were directly associated with working memory deficits.

Keywords: GWAS; child development; cognition; obesity; polygenic risk score.

FIGURE 1

Mediation analysis of Adulthood-BMI-PRS, Childhood-BMI, and Childhood-Working memory. The solid line indicates the path that was statistically significant and the dashed lines indicate those that were estimated but not statistically significant. BMI, body mass index; PRS, polygenic risk score. ^∗∗p < 0.01.