Caffeine's dopamine dependent aversive depend on the dopamine D₂ and D₁ receptor, however the D₂ receptor is not involved in the locomotor stimulant effect of caffeine.
(A) C57BL/6 mice conditioned with 10 mg/kg caffeine showed significant aversions while D₁KO mice and D₂KO mice did not. D₂KO mice demonstrated preferences for this dose of caffeine, while D1KO did not show an aversion or a preference. Bars represent mean difference scores ± S.E.M. * indicates P < 0.05 for comparison of difference scores to zero that were corrected for multiple t tests.
(B) The dopamine D₁ receptor does not mediate the rewarding effects of caffeine. D₂KO mice show a conditioned place preference to 10 mg/kg caffeine, which their wild type littermates find aversive. When D₂KO mice were pretreated with the dopamine D₁ receptor antagonist SCH23390 their preferences to caffeine did not diminish, indicating the D₁ receptor is not the substrate of caffeine reward. Bars represent mean difference scores and error bars are SEM. NS indicates no main effect of pretreatment (P > 0.05) and * indicates a main effect of genotype (P < 0.05).
(C) A non-dopaminergic substrate mediates the aversive effects of 100 mg/kg caffeine. D₁KO mice and D₂KO mice show conditioned place aversions equivalent to those seen in wild type littermates when given 100 mg/kg caffeine despite the fact that D₁KO mice did not show a preference or an aversion and D₂KO mice show conditioned place preferences to 10 mg/kg caffeine. Bars represent mean difference scores and error bars are SEM. NS indicates no main effect of genotype on difference scores.
(D) C57BL/6, D₁KO and D₂KO mice all showed equivalent locomotor stimulation after 10 mg/kg i.p. caffeine. Locomotor activity in D₂KO mice was significantly lower than locomotor activity in D1KO mice. Bars represent mean locomotor counts for saline and caffeine trials ± S.E.M. * indicates a significant difference between activity of D1KO and D2KO mice (P < 0.05).

The adenosine A₁ receptor is not critical for the stimulant effects of caffeine and is not involved in the aversive or rewarding effects of caffeine either alone or in combination with the A_2A receptor.
(A) A₁KO mice showed aversions to 10 mg/kg caffeine that did not differ from those seen in wild type littermates. Knockout mice also showed preferences to 10 mg/kg caffeine when pretreated with 0.8 mg/kg alpha-flupenthixol and preferences did not differ from those seen in wild type littermates. Bars represent mean difference scores ± S.E.M.
(B) A₁KO mice show locomotor stimulation in response to caffeine equivalent to locomotor stimulation seen in wild type littermates. Bars represent mean locomotor counts ± S.E.M. * indicates a significant difference between saline and caffeine locomotor scores in A₁KO mice (P < 0.05) but not in wild type littermates (P > 0.05).
(C) dKO mice show aversions to 10 mg/kg caffeine and preferences to 10 mg/kg caffeine when pretreated with 0.8 mg/kg alpha-flupenthixol that do not differ significantly from those seen in C57BL/6 mice. Bars represent mean difference scores ± S.E.M.

Caffeine has dopamine independent rewarding effects and dopamine-dependent aversive effects that are mediated in the central nervous system.
(A) Non-pretreated mice showed significant aversions to caffeine at doses of 10, 15, 30, and 100 mg/kg, respectively. Animals pretreated with 0.8 mg/kg alpha-flupenthixol showed significant preferences for caffeine at doses of 3, 10, 15, and 30 mg/kg. Open markers represent non-pretreated animals and filled markers represent alpha-flupenthixol pretreated groups. Data points represent mean difference scores ± S.E.M. * indicates P < 0.05 for comparison of difference scores to zero.
(B) Caffeine's aversive effects are centrally-mediated. C57BL/6 mice treated i.p. with 8-SPT, a caffeine analogue unable to cross the blood-brain barrier, did not show conditioned place aversions to the drug. Bars represent the mean amount of time spent in the drug-paired and the saline-paired environments on test day ± S.E.M. NS indicates P > 0.05.

The adenosine A_2A receptor is not directly involved in the aversive or rewarding effects of caffeine but is critical for caffeine's locomotor stimulant effects.
(A) Both C57BL/6 and A_2AKO mice showed similar aversions to 10 mg/kg i.p. caffeine. Pretreating C57BL/6 and A_2AKO mice with 2.0mg/kg alpha-flupenthixol blocked the aversive effects of caffeine and revealed preferences in both genotypes. Bars represent mean difference scores ± S.E.M. Non-pretreated and pretreated A_2AKO groups did not differ from their respective C57BL/6 control groups.
(B) Both C57BL/6 and A_2AKO mice showed similar aversions to 15 mg/kg i.p. caffeine. Pretreating C57BL/6, but not A_2AKO mice, with 0.8 mg/kg alpha-flupenthixol blocked the aversive effects of caffeine and revealed a preference for caffeine. A_2AKO mice pretreated with 0.8 mg/kg alpha-flupenthixol showed an aversion to caffeine that differed from the response seen in pretreated C57BL/6 mice. Bars represent mean difference scores ± S.E.M. * indicates P < 0.05 for comparison of neuroleptic-pretreated C57BL/6 mice to pretreated A_2AKO mice.
(C) C57BL/6 but not A_2AKO mice showed locomotor stimulation when administered 10 mg/kg caffeine i.p., however both genotypes showed a locomotor response to the 15 mg/kg dose of caffeine. There was a significant effect of drug treatment and genotype (both P < 0.05). Bars represent mean locomotor counts for saline and caffeine trials ± S.E.M. stats
(D) Caffeine partially counteracts the locomotor depression induced by alpha-flupenthixol in C57BL/6 but not A_2AKO mice. Both C57BL/6 mice and A_2AKO mice experienced locomotor depression relative to baseline (dotted line) when treated with 0.8 mg/kg alpha-flupenthixol. This depression was partially counteracted by treatment with 15 mg/kg caffeine in C57BL/6 but not A_2AKO mice. Bars represent locomotor scores when mice were given caffeine or saline after pretreatment with alpha-flupenthixol as a percentage of locomotor scores in the absence of alpha-flupenthixol. The dotted line indicates baseline locomotor activity of each genotype in the absence of neuroleptic.

Lesions of the tegmental pedunculopontine nucleus in mice do not affect aversions or dopamine blockade-induced conditioned place preferences to i.p. caffeine.
(A) Both sham TPP-lesioned and TPP-lesioned mice showed aversions to 10 mg/kg i.p. caffeine. Pretreating sham TPP-lesioned and TPP-lesioned mice blocked the aversive effects of 10 mg/kg i.p. caffeine in both groups and revealed a significant preference. Bars represent mean difference scores ± S.E.M. Aversions and preferences of lesioned groups did not differ significantly from sham-lesioned controls.
(B) Coronal section of the brain; the grey rectangle indicates the region depicted in photographs and dotted lines indicate the approximate location of the tegmental pedunculopontine nucleus. Photographs on the right depict a representative bilateral sham TPP lesion (top) and an excitotoxic TPP lesion (bottom). Arrows indicate gliosis in the region of the TPP.