Zidovudine (AZT) is an effective treatment in preventing perinatal transmission of HIV-1; however, a continuous re-evaluation of the risk-benefit ratio of human exposure to this drug is suggested by both clinical and animal studies. The objective of this study was to assess the medium and long-term effects of pre-postnatal AZT treatment on mouse social and emotional behaviour and the possible interactions between AZT exposure and disruptions in the mother-infant relationship. Pregnant CD-1 mice were administered per os with AZT (160 mg/kg) from pregnancy day 10, throughout delivery, to lactation day 10. In half of the litters, the offspring was separated from the mother for 3 h from postnatal days 2 (PND2) to PND14. On PND35, a 30-min social interaction test was performed and corticosterone levels were measured at the end of the session. On PND80, long-term effects of AZT on emotionality were assess by means of an elevated plus-maze. Results indicate that, on PND35, previous AZT exposure affected social behaviour of the experimental subjects, reducing aggressive interactions in males, while decreasing investigative behaviours in females. At adulthood, AZT inhibited exploratory behaviour in the plus-maze while increasing the frequency of risk-assessment postures in male mice. As for maternal deprivation, this early manipulation exerted a pro-aggressive effect in adolescent male mice, deprived subjects being overall characterised by higher activity levels and a deficit in habituation, an effect also observed in the plus-maze. A significant interaction between AZT and maternal deprivation was found for affiliative behaviours. As for corticosterone levels, no AZT effect was found, while maternal deprivation tended to reduce elevations of this hormone in response to stressful stimuli. Overall results from this study indicate that both AZT exposure and maternal deprivation induced gender-dependent changes in social and emotional behaviour both during adolescence and at adulthood.