This study provides a demonstration in the rat of a clear genetic difference in the propensity for addiction-related behaviors following prolonged cocaine self-administration. It relies on the use of selectively bred high-responder (bHR) and low-responder (bLR) rat lines that differ in several characteristics associated with "temperament," including novelty-induced locomotion and impulsivity. We show that bHR rats exhibit behaviors reminiscent of human addiction, including persistent cocaine-seeking and increased reinstatement of cocaine seeking. To uncover potential underlying mechanisms of this differential vulnerability, we focused on the core of the nucleus accumbens and examined expression and epigenetic regulation of two transcripts previously implicated in bHR/ bLR differences: fibroblast growth factor (FGF2) and the dopamine D2 receptor (D2). Relative to bHRs, bLRs had lower FGF2 mRNA levels and increased association of a repressive mark on histones (H3K9me3) at the FGF2 promoter. These differences were apparent under basal conditions and persisted even following prolonged cocaine self-administration. In contrast, bHRs had lower D2 mRNA under basal conditions, with greater association of H3K9me3 at the D2 promoter and these differences were no longer apparent following prolonged cocaine self-administration. Correlational analyses indicate that the association of H3K9me3 at D2 may be a critical substrate underlying the propensity to relapse. These findings suggest that low D2 mRNA levels in the nucleus accumbens core, likely mediated via epigenetic modifications, may render individuals more susceptible to cocaine addiction. In contrast, low FGF2 levels, which appear immutable even following prolonged cocaine exposure, may serve as a protective factor. addiction | dopamine | fibroblast growth factor | nucleus accumbens | reinstatement A pproximately 16% of adults in the United States report drug use within the past year (1). However, not everyone who experiments with drugs becomes an addict, as an estimated 8.5% of the population, or 25 million Americans, meet Diagnostic and Statistical Manual of Mental Disorders IV (2) criteria for substance abuse and dependence (1). Environmental and societal factors play a role in addiction liability (e.g., refs. 3-5), and there is ample evidence demonstrating a role for genetic factors (e.g., refs. 6-10). However, studying the interplay among these factors is difficult in human studies because of the inability to control for environmental factors and the challenge of parsing causes from consequences. Preclinical animal models are therefore essential for defining the complex interactions between genes and environment, and uncovering the neural mechanisms that might render an individual more susceptible to drug addiction. The first animal model characterizing individual differences in the propensity to take drugs of abuse was introduced over two decades ago by Piazza et al. (11), who showed that, like humans, only some rats readily self-administer such drugs. Furthermore...