Controversies in translational research: drug self-administration

Laboratory animal and human models of drug self-administration are used to evaluate potential pharmacotherapies for drug abuse, yet the utility of these models in predicting clinically useful medications is variable.

The objective of this study was to track how antagonist, agonist, and partial agonist medication approaches influence heroin and cocaine self-administration by rodents, non-human primates, and humans and to compare these results to clinical outcomes.

Across species, heroin self-administration was decreased by all three medication approaches, paralleling their demonstrated clinical utility. The heroin data emphasize the importance of assessing a medication's abuse liability preclinically to predict medication abuse and compliance and of considering subject characteristics (e.g., opioid dependence) when interpreting medication effects. For cocaine, the effects of ecopipam, modafinil, and aripiprazole were consistent in the laboratory and clinic, provided that the medications were administered repeatedly before self-administration sessions. Modafinil attenuated cocaine's reinforcing effects in the human laboratory and improved treatment outcome, while ecopipam and aripiprazole increased the reinforcing effects of cocaine and do not appear promising in the clinic.

The self-administration model has reliably identified medications to treat opioid dependence, and the recent data with modafinil suggest that the human laboratory model also identifies medications to treat cocaine dependence. There have been numerous false positives when subjective effects are the primary outcome measure, but not when self-administration is the outcome. Factors relevant to the predictive validity of self-administration procedures include medication maintenance and the concurrent assessment of a range of behaviors to determine abuse liability and the specificity of effect.