Genetic, prenatal, postnatal, and endocrine influences on adolescent smoking

Multiple bio-behavioral mechanisms of the development of adolescent smoking have been identified, including genetic risk, prenatal adversity, blunted cortisol reactivity, less warm/more hostile parenting, and behavior problems... [ view full abstract ]

Multiple bio-behavioral mechanisms of the development of adolescent smoking have been identified, including genetic risk, prenatal adversity, blunted cortisol reactivity, less warm/more hostile parenting, and behavior problems within the broader developmental context. Each developmental influence highlighted here does not occur in a vacuum, and what is needed are studies that examine these multiple influences together to identify the strongest predictors and developmental pathways to substance use. The genetic underpinnings of cortisol functioning are relatively well documented, but few studies have tested whether genetic or environmental mechanisms (endogenous or exogenous) underlie cortisol-smoking associations, despite evidence that some of the same genes are associated with cortisol and smoking (in different studies). We tested whether a polygenic score of cortisol-related genes, prenatal adversity, and parenting at age 11 years predicted cortisol reactivity to a social stress challenge at age 11 years, trajectories of internalizing and externalizing problems from age 11-16, and smoking frequency at 16 years using Genome-wide association and phenotypic data from 2230 Dutch adolescents in the TRacking Adolescents’ Individual Lives Survey. We used a 2-fold (discovery/test) set approach to (1) create polygenic scores related to cortisol function and then (2) test the main effects and full structural equation models including genetic, prenatal, parenting, and behavioral trajectory predictors of smoking. Polygenic scores were constructed from genes empirically linked to cortisol functioning and/or listed in one of several biological pathways key to cortisol production: 217 genes were identified, 1071 (non-imputed) single-nucleotide polymorphisms (SNPs) on 1354 individuals were included after standard quality control and pruning for linkage disequilibrium > .7. Zero-order correlations revealed that no SNPs were significantly related to cortisol after adjustment for multiple testing. Results generally suggested that cortisol, prenatal adversity, and colder parenting all predicted smoking frequency, but that the polygenic score was unrelated to a) cortisol, b) other predictors, and c) smoking frequency. Findings suggest that cortisol relates to smoking via behavioral, and not via genetic mechanisms.