Abstract

Abundant evidence from family, adoption, and twin studies point to large genetic contributions to individual differences in vulnerability to develop dependence on addictive substances. Twin data suggest that much of this genetic vulnerability is shared by individuals who are dependent on a variety of addictive substances, though some are likely to be substance-specific. Twin data for abilities to quit smoking provide some of the best evidence for genetic influences on abilities to achieve and maintain abstinence from use of an addictive substance.

Here, we approach genetic contributions to addiction vulnerability and ability to quit from the bottom-up. We describe a series of working hypotheses concerning the genetic architecture of addiction that are supported by evidence of differing strengths. These data, overall, provide a substrate to improve understanding of substance dependence and the ability to quit smoking. With better understanding of genetic influences on these phenotypes, we will be better positioned to improve understanding of the large environmental influences on these phenotypes, to personalize treatments, and to personalize prevention strategies for individuals at especial risk.

Keywords

Drug Abuse; Substance Abuse; Addiction; Gene; DNA Variants

1 Introduction

The sizable genetic influences on individual differences in vulnerability to developing dependence on an addictive substance (see below) are likely to overlap substantially with overall genetic influences on abilities to quit, as well as with genetic influences on a number of related phenotypes and comorbid conditions. At a molecular level, this means that DNA sequence variants that are passed from generation to generation can, in permissive environments, work to alter these vulnerabilities. We can thus start from an overview of the types of DNA variants that are likely to contribute to these vulnerabilities. We include novel data for possible roles for copy number variants (CNVs). We then move upscale to discussion of shared genetic influences on comorbidities and on the possible implications for thinking about, treating, and preventing addictions.

2.1 Working Hypothesis I: Genomic Variants of Several Classes and Differing Frequencies Contribute to Vulnerability to Addiction and Ability to Quit

Thinking about the genetic architectures of vulnerability to addiction and ability to quit can start from the microscopic level: What sorts of variants in the genome are likely to confer these risks? What are the ways in which these variants are distributed in individuals, families, and populations? (See Table 1)

Neither linkage studies, which assess the way in which DNA markers and disease vulnerability move together in families, nor most association studies, which assess the ways in which single nucleotide polymorphism (SNP) DNA markers and disease are associated in nominally unrelated individuals, provide direct assessments of the actual DNA variants that contribute to linkage or association findings. Follow-up studies from these findings, and initial data from studies of CNV probes, are beginning to paint a rich picture of the ways in which specific variants may influence vulnerability to dependence. Data from exome or genome-wide resequencing will soon provide a deluge of variants that have the potential for association with addiction in at least some individuals. Separating epiphenomena from truly associated variants from this work will be likely to take large amounts of effort over many years, however.

A common Mendelian disorder, cystic fibrosis, provides an analogy for the types of DNA variants that are likely to influence addiction vulnerability and ability to quit. Almost 2000 separate DNA variants that migrate in families along with cystic fibrosis are described: http://projects.tcag.ca/variation/. These include variants that influence the sequence of the cystic fibrosis transmembrane regulator (CFTR) protein in which variation causes cystic fibrosis by changing amino acids or terminating the protein prematurely. These disease-associated variants also provide a variety of different influences in intronic, 5′flanking and even 3′ flanking regions of the CFTR gene. Further, the entire constellation of CFTR variants and the disease are more abundant in individuals with European genetic backgrounds. Specific variants are much more prominent in individuals from specific regions of Europe. This degree of complexity for even a relatively common disorder with classical Mendelian inheritance patterns presages substantial molecular genetic complexity for addiction.

Some of the most detailed information about individual addiction-associated variants comes from studies of candidate genes. Association with levels of expression and with missense variants that alter receptor function has been identified for genes in the chromosome 15 nicotinic acetylcholine receptor gene cluster that are associated with measures of smoking quantity/frequency, for example (Wang et al., 2009). Some data indicate subtle differences in the binding properties associated with OPRM1 missense variants that alter its N-terminus and have been associated with differential efficacies of opiate antagonist therapies for alcohol in at least some studies (Bond et al., 1998; Oslin et al., 2003). We have identified splicing variants in NRXN3 that are associated with addiction vulnerability (Hishimoto et al., 2007); these findings have been recently supported by initial studies of mice engineered to differentially express these splicing variants (Boucard et al., 2011). It appears safe to postulate that all types of DNA variants will eventually be found to contribute to addiction vulnerability and abilities to quit.

We specifically sought evidence for overrepresentation of CNVs in Affymetrix 6.0 genotyping studies of DNAs pooled from groups of 20 dependent and 20 control individuals recruited in Baltimore, MD (Drgon et al., 2009a). Validating studies for use of the CNV probes in pooled samples provided evidence for poorer overall performance of some of the CNV probes than of the SNP probes assessed on the same arrays (Figure 1). Nevertheless, principal components analyses of data from all CNV probes provided interesting components that are strongly associated with addiction phenotypes (Figures 2 and 3).

This data provide some of the only evidence of which we are aware that CNVs are enriched, as an overall class, in dependence. CNVs that were identified in both independent African- and European-American samples that compared individuals with (1) substantial use of and, almost always, dependence on at least one illegal substance to (2) controls with no significant lifetime use of any addictive substance are listed in Table 2.

Studies of the frequency of variants associated with dependence can also be considered in light of recent evidence for the extent to which the same genomic regions are identified in independent addiction vulnerability samples of the same vs different racial/ethnic backgrounds (Drgon et al., 2011; Johnson et al., 2011). We and others have identified association signals in samples from both African- and European-American genetic backgrounds. A single quit-success genotype scores can differ between quitters and nonquitters from both European- and African-American genetic backgrounds (Uhl et al., 2010a). Nevertheless, when we have recently compared addiction associations in independent samples of European or African genetic backgrounds, there are much stronger overlaps between different studies of European samples than there are in comparison of European vs African samples (Drgon et al., 2011; Johnson et al., 2011). Taken...