Our multivariate GWAS (total n=1,025,550) revealed that a general liability to addiction to alcohol, tobacco, cannabis, and opioids is characterized by a highly polygenic architecture. The addiction-rf correlated with other medically relevant traits and illnesses. Substance-specific genetic factors with relatively less polygenic architectures, and in some instances, a single SNP of larger effect (e.g., ADH1B for alcohol) were also identified. Outlining this common genetic factor - addiction-rf - suggests that current substance-specific diagnostic classification schemes (e.g., DSM) only partially represent the phenomenon of SUD at the genetic level of analysis.

We found 17 genomic loci significantly associated with addiction-rf, and 27 substance-specific loci. Post-hoc fine-mapping, annotation, and exploratory drug repurposing analyses highlight the potential therapeutic relevance of the discovered loci. The addiction-rf PRS was associated with many medical conditions characterized by high morbidity and mortality rates, including psychiatric illnesses, self-harming behaviors, and somatic diseases that could be consequences of chronic substance use (e.g., chronic airway obstruction) or precursors to heavy substance use (e.g., chronic pain). Finally, in a sample of drug-naïve children, the addiction-rf PRS was correlated with impulsivity and externalizing behaviors, a family history of hospitalization for and diagnosis of severe mental illness, thought problems, and neighborhood deprivation, which reflects the shared genetic and gene-environment correlations with phenotypes and exposures present early in life.

Our analyses suggest that the regulation/modulation of dopaminergic genes, rather than variation in dopaminergic genes themselves, is central to general addiction liability. DRD2 was the top gene signal, which was mapped via chromatin refolding, suggesting a regulatory mechanism. The role of striatal dopamine in positive drug reinforcement is well established³². DRD2 plays a role in reward sensitivity and may also be central to executive functioning³³ – the interplay of reward and cognition is likely relevant throughout the course of addiction. These complementary observations reinforce the role of dopamine signaling in addiction^34,29. Our results give some evidence that chromatin refolding plays a role in these processes.

Other regulatory effects on dopaminergic pathways were supported by the signal at PDE4B, which has been implicated in prior GWASs of disinhibition traits³⁵. The PDE4 antagonist, ibudilast, has been shown to reduce heavy drinking among patients with AUD^36,37 and also shown to reduce inflammation in methamphetamine use disorder³⁸. The phosphodiesterase (PDE) system has been proposed as a dopaminergic regulation mechanism, and PDE inhibitors have been tested as treatments for cognitive impairments in schizophrenia that arise from antipsychotic antagonism of dopamine receptors, albeit with side effects³⁹. Further, animal studies suggest that the PDE system is associated with down-regulation of drug-seeking behaviors across opioids, alcohol, and psychostimulants⁴⁰, convergent evidence of the potential translational significance of this gene system for therapeutics⁴¹.

The addiction-rf PRS was associated with general and specific SUD liability in an independent sample. The addiction-rf PRS predicted ∼6% of OUD variance, which is nearly half the total SNP-heritability of OUD¹⁶. The addiction-rf PRS also predicted variance in CoUD - as CoUD was not included in the development of the addiction-rf (lack of well-powered CoUD GWAS), these findings highlight the generalizability of the addiction-rf beyond alcohol, tobacco, cannabis and opioids. Further, the addiction-rf PRS explained appreciable variance for a phenotypic factor comprised of AUD, CUD, OUD, TD and CoUD (R² = 6.6%), suggesting that combining information across SUDs may be a useful strategy in PRS analyses of addiction.

Considering how readily many of the top hits and mechanisms can be translated into medication targets (e.g., PDE4B), we conducted a reverse transcriptional drug repurposing analysis of addiction-rf. Rather than focusing on individual gene targets, this method uses a transcriptional profile to predict the top drugs that reverse the transcriptional pattern that we derived from a transcriptome-wide association analysis. We found >100 potential perturbagens that partially reverse the transcriptional patterns underlying general addiction risk (Supplemental Table 17). Some of these drugs have existing evidence for treatment of SUDs, such as varenicline, an efficacious smoking cessation drug⁴² that may also reduce craving for alcohol in smokers and drinking in non-smokers⁴³. Other drugs identified in the repurposing analysis that are being actively investigated include mifepristone, which is currently under clinical investigation for the treatment of alcohol dependence^44,45; riluzole, for which there is early evidence of a treatment effect in stimulant dependence^46,47; and reboxetine, that may have utility in the treatment of CoUD⁴⁸.

Substance-specific genetic signals fell primarily into three broad categories: drug-specific metabolism (e.g., ADH1B for PAU), drug receptors (e.g., CHRNA5 for PTU, OPRM1 for OUD), and general neurotransmitter mechanisms (e.g., CACNA1A for CUD). In particular, alcohol-specific effects associated with ADH1B and other alcohol dehydrogenase genes⁴⁹ and tobacco-specific effects associated with CHRNA5 and acetylcholine receptor genes were an order of magnitude greater than those noted for the addiction-rf. Surprisingly, even after accounting for the addiction-rf, dopaminergic genes (DBH and PDE1C in particular) were implicated in substance-specific effects for tobacco (PTU). On the other hand, CUD-specific genes did not include well-studied receptor targets (e.g., CNR1) or metabolic mechanisms (e.g., Cytochrome P450 genes). However, the cannabis-specific loci were eQTLs for CHRNA2, and there is some evidence that CNR1 and CHRNA2 expression are anti-correlated in some brain regions⁵⁰. It is also likely, given the high loading (0.74) of CUD on addiction-rf, that current GWASs of CUD capture genetic vulnerability that is relatively non-specific. Alternatively, effect sizes for cannabis-specific variants may be too small to detect in the available data.

Genome-wide genetic correlations and PheWAS across multiple different samples supported the role of pleiotropy and/or causality in the relationship between problematic substance use and other disinhibitory behaviors across the lifespan. Age at first sexual intercourse was amongst the phenotypes most highly genetically correlated with addiction-rf. Even in predominantly substance-naïve youth, the addiction-rf PRS was associated with externalizing behaviors, which reinforces the developmental role of these behaviors in the etiology of drug use. Despite the genetic overlap between the addiction-rf and a recent index of externalizing behaviors (rG=0.63)³⁵, a significant portion of the variance in addiction-rf was distinct.

Our analyses highlight the robust genetic association of addiction-rf with serious mental and somatic illness. The addiction-rf PRS was more strongly associated with using drugs to cope with internalizing disorder symptoms (anxiety, depression; rg=0.60-0.62) than with the individual psychiatric traits and disorders themselves (rg=0.3), suggesting that genetic correlations between SUDs and mood disorders may partially be attributable to a predisposition to use substances to alleviate negative mood states (“self-medication”)⁵¹. Self-harm as indexed by the ingestion of medication, suicide attempt, and family history of suicide attempt were also highly correlated with the addiction-rf (rg range: 0.45-0.62). The findings regarding personal and family history of suicidal thoughts and behaviors are consistent with other recent reports that addiction is more highly correlated with suicidality than most other behaviors^52-54. As with internalizing disorders, although the observed genetic correlation could reflect a causal pathway from SUD to suicidal thoughts and behaviors⁵⁵, latent causal variable analysis revealed a non-significant causal effect (p=0.72, Supplemental Table 18).

The PheWAS also provided insight into potentially complex mechanisms of genetic liability to environmental pathways of risk. In addition to indices of socio-economic status (SES), the addiction-rf was correlated with maternal smoking during pregnancy and ADHD, in line with evidence that effects ascribed to the prenatal environment may also be mediated by the inheritance of risk loci^56,57. The addiction-rf PRS was associated with a family history of serious mental illness, which likely represents an amalgam of genetic and environmental vulnerability⁵⁸. Finally, disability and SES were also associated with polygenic risk, further supporting the association between environmental risk factors and common genetic effects on SUD liability^9,59,60.

This study has limitations. First, our GWAS in individuals of African-ancestry had few discoveries, underscoring the need for systematic data collection on SUDs in globally representative populations. Still, we chose to analyze and present these data as their exclusion only furthers disparities in genetic discoveries. Second, although we discovered many loci, they accounted for only a small proportion of the total variance. More samples, particularly those in diverse populations, and the integration of rarer variants are needed to discover the biological pathways that fall below genome-wide significance or are missed in GWAS. Finally, the authors note that despite interesting correlations between the addiction-rf PRS and a variety of traits, effect sizes were not of a clinically-relevant magnitude nor were analyses designed to make inferences that could apply in a clinical setting. Furthermore, some of the observed associations (e.g., in the ABCD® study where we examine a host of parental characteristics using offspring PRS) are likely confounded by other factors (e.g., SES), which are beyond the scope of detailed study here, and hence have no application in clinical prognostication. Critically, our analyses clearly demonstrate that genetic risk to addictions is polygenic and comprised of very small effects of many loci, and that this genetic vulnerability is accompanied by environmental risk and resilience factors. Therefore, utilizing these PRS alone to forecast outcomes in developing youth is unlikely to be of utility in the absence of rigorously conducted follow-up studies. Consequently, the authors strongly caution against using the findings from this study, including the PRS, beyond research, and particularly across ancestral groups where they can generate spurious findings if not appropriately applied.