Genetic influences on substance use behaviour and disorders and psychiatric disorders

Abstract

Substance use, substance use disorders (SUDs), and psychiatric disorders are prevalent, comorbid and heritable. Substance use is common in people with psychiatric disorders, psychological distress, or somatic distress. These conditions may be further complicated by the tendency to use multiple substances. I hypothesised that shared genetic risk underlay these co-occurring conditions. I tested this hypothesis by investigating the common and specific genetic influences underlying the use of different substances, SUDs, and some common psychiatric disorders.Chapter 1 will give a background for the phenotypes and statistical methods used in the four results chapters.Chapter 2 estimated the genetic and environmental influence on the variation of four groups of disorders, including psychological distress, somatic distress, affective disorders, and substance use, by conducting twin modelling (N=1548~ 2132 twin individuals). Firstly, I modelled the four disorder groups separately with item response theory (IRT) to create an IRT theta score as a refined phenotype for each of them. These IRT scores and other relevant phenotypes were then analysed with univariate and multivariate twin modelling. I provided the first evidence that the psychological distress and somatic distress were heritable and that the covariation between these two subscales was largely explained by common genetic factors. In addition, I identified an independent pathway model that best explained the covariation between the four IRT scores. This model showed the covariation was largely explained by environmental factors (53%~ 64%) that were specific to each of the four phenotypes. Genetic factors common to the four phenotypes explained a small but non-negligible portion (3%~22%) of the covariation.Chapter 3 tested the associations between licit substance use and illicit substance use. I first calculated polygenic risk scores (PRS) using genome-wide association study (GWAS) meta-analysis summary statistics derived from a large discovery sample GWAS & Sequencing Consortium of Alcohol and Nicotine use (GSCAN; 258,797- 632,802 subjects) for four smoking and one alcohol phenotypes: regular smoking initiation (PRS-SI), age of initiation of regular smoking (PRS-AI), cigarettes per day (PRS-CPD), smoking cessation (PRS-SC), and drinks per week (PRS-DPW). These PRSs were then used to predict individual differences in a total of 22 substance use phenotypes in an independent target sample of 2,463 young Australian adults. PRS-SI explained variation in the liability of initiating cocaine (0.67%), amphetamine (1.54%), hallucinogens (0.72%), ecstasy (1.66%) and cannabis initiation (0.97%), as well as DSM-5 alcohol use disorder (0.72%). PRS-DPW explained 0.75%, 0.59% and 0.90% of the variation of initiating cocaine, amphetamine and ecstasy use respectively. This study provided the first evidence for genetic overlap between the polygenic risks for smoking or alcohol, and the risk of initiating major classes of illicit substances.Chapter 4 tested the associations between the same set of the five PRS used in Chapter 3 and individual differences in 15 licit substance use and common psychiatric disorders in 13,999 Australian adults. Ten out of 75 pairwise-combinations of discovery and target phenotypes remained significant after correcting for multiple testing. PRS-SI was positively associated in the independent target sample with smoking initiation, drinks per week, conduct disorder, and nicotine dependence defined by DSM-IV and FTND (R2 range: 1.98% ~ 5.09%). PRS-AI was negatively associated with DPW (R2: 3.91%). PRS-CPD was positively associated with DSM-IV nicotine dependence and conduct disorder (R2: 1.56% ~ 1.77%). PRS-DPW was positively associated with DPW only (R2: 3.39% ~ 6.26%). A significant interaction between PRS and sex was detected in the associations between DPW and PRS-SI, PRS-AI or PRS-DPW. I provided the first evidence for the association between polygenic risk for smoking and risk for conduct disorder.Chapter 5 investigated the phenotypic associations, genetic correlations and causal associations between the initiation or use of four common substances alcohol, caffeine, cannabis, and nicotine. GWAS summary results for ten traits were derived or obtained from three large samples GSCAN, UK Biobank (UKB), and International Cannabis Consortium (ICC). After correcting for multiple testing, 33 of the 45 trait pairs were genetically correlated. Results from two-sample Mendelian randomisation (MR) estimators showed that most exposure and outcome traits were not causally associated. However, a causal relationship was found between regular smoking initiation and caffeine consumption, with two-fold increased prevalence of regular smoking initiation causally associated with an increase of 40.17 mg caffeine consumed. This association was robustly resistant to horizontal pleiotropy and SNP outliers.In conclusion, I showed that the association between psychological distress and somatic distress was largely due to genetic covariance. I identified shared genetic risk between the use of tobacco or alcohol, and the risk for initiating five different illicit substances, nicotine dependence, and conduct disorder. Using various MR approaches, I provided evidence for a positive causal effect of regular smoking initiation on caffeine consumption. These findings suggest that substance-substance or substance-disorder associations were partially driven by shared genetic aetiology, but I found little evidence for causality in most of the exposures and outcomes.