There is evidence that both particular genes (heritability as high as 60%) and particular environments contribute to vulnerability to subsequent alcohol abuse and dependence [Mcgue, 1999; Enoch and Goldman, 2001; Lesch 2005]. However, there is relatively little data on gene-environment interactions in the aetiology of alcohol use disorders (AUDs) [van der Zwaluw and Engels, 2009]. Further, little is known about the role of internalising behaviour in mediating the relationship between genes, environments and subsequent AUDs.

The Avon Longitudinal Study of Parents and Children (ALSPAC) provide a unique opportunity to try and determine the relationship between genes and environments in AUDs, through the mediation of internalising behaviour. First, ALSPAC includes data on genetic variants and on adverse environments (e.g. stressful life events (SLEs)). Second, ALSPAC includes comprehensive data on internalising behaviour experienced during childhood.

Heavy alcohol use is positively associated with the experience of environmental stressors such as job, health-related, social and legal stress [Dawson et al., 2005]. In turn, the behavioural response to stress is mediated by corticotrophin-releasing factor (CRF) [Arborelius et al., 1999; Binder and Nemeroff, 2010]. Studies involving animal models have shown that polymorphisms within the CRF receptor gene (CRHR1), together with the experience of environmental stress factors, result in greater alcohol consumption [Sillaber et al., 2001; Hansson et al., 2006]. Similarly, in humans, polymorphisms in the CRHR1 gene interact with negative life events to predict heavy alcohol use [Treutlein et al., 2006; Blomeyer et al., 2008]. Besides the CRHR1 gene, the CRF pathway consists of the following genes: POMC, UCN, CRH, CRHR2, UCN3 and UCN2 [PharmGkB. Date accessed: 13 March 2013]. These genes are relatively unexplored in terms of vulnerability to alcohol dependence.

It is often argued that AUDs are characterised by externalising behaviour [Kumpulainen, 2000; Englund et al., 2008]. Findings regarding the link between behavioural inhibition/internalizing symptoms and AUDs are less emphasised in the literature, although a few studies have shown that internalising symptoms, such as anxiety and depression, are associated with alcohol use [Loeber et al., 1999; Fite et al., 2006]. CRF has also been shown to play a role in internalising behaviours; indeed polymorphisms within CRHR1 have shown to have a significant association with anxious temperament and have an influence on the metabolic activity of local brain regions in rhesus macaques [Rogers et al., 2012]. There has been little work, however, on the relationship between variants in the CRF pathway and internalising behaviour in humans.

Methodology

1) Select variants from the genes involved in the CRF pathway based on a systematic review of the literature.

2) Include genotype in the models as a (a) quasi-continuous, (b) time-invariant predictor of latent continuous and/or (c) categorical variable

i. Test for association with the extended alcohol use phenotype measured by the 10-item Alcohol Use Disorders Identification Test (AUDIT) administered at 16 years of age or later [Heron et al., 2012].

3) Identify mediating pathways by which genes confer susceptibility

i. Hypothesized that CRF genes will be likely to confer risk via internalising behaviour evident in early childhood. Internalising behaviour will be measured by the parent-rated, emotional subscale of the Strength and Difficulties Questionnaire (SDQ), administered at 7 years of age or older as described by Araya et al. (2008) and Evans et al. (2008).

ii. Hypothesized that CRF genes will be related directly to alcohol use and risk pathways in adolescence that emerge concomitantly with AU and not involved in pathways with origins in childhood

4) Test GxE with environments previously shown to moderate genetic risk. Environmental factors such as SLEs will be measured using Section D of the parent-rated questionnaire "My Son's/Daughter's Health and Behaviour-42 months". SLEs will also be assessed by the "Life Events" questionnaire completed by the Mother of the "child" at 18, 30, 42 and 57 months as well as 5, 6 and 8 years and completed by the "child" at 16 years of age.

i. Regressing outcome variables onto (a) genotype (b) an environment, and (3) and interaction term reflecting the product of genotype and environment.