Expression of stable and reliable preference and aversion phenotypes following place conditioning with psychostimulants

Rationale and objectives: Drug-seeking behavior occurs more readily in some individuals than others. This phenomenon is considered in studies of drug self-administration in which high drug-seeking/taking individuals can be identified. In contrast, studies of conditioned place preference (CPP) often involve a random sample of drug-naïve rodents that includes phenotypes not considered relevant to addiction. The main objective of the current studies was to determine if a priori identification of different conditioning phenotypes could improve the validity and sensitivity of CPP expression as a preclinical test for vulnerability to addiction. Methods and results: Analysis of cocaine place conditioning data from 443 Swiss-Webster mice revealed a trimodal distribution with peaks corresponding to means of k = 3 clusters. The cluster means occurred at high, low, or negative preference scores, the latter suggesting a phenotype acquiring conditioned place aversion (CPA). The same clusters were identified in mice conditioned with methamphetamine, MDPV, or amphetamine, and these clusters remained stable and reliable during three additional expression tests spaced at 24 h. A meta-analysis of effect sizes obtained from CPP literature revealed a positively skewed distribution affected by sample size, consistent with the existence of a CPA phenotype within the populations tested. A dopamine receptor antagonist, flupentixol, blocked cocaine CPP expression in a group containing all phenotypes, but sensitivity improved markedly when CPA phenotypes were excluded from the dataset. Conclusions: These studies suggest that taking phenotype into consideration when designing place conditioning studies will improve their application as a preclinical tool in addiction biology and drug discovery.