According to the adenosine hypothesis of schizophrenia, the classically associated hyperdopaminergic state may be secondary to a loss of function of the adenosinergic system. Such a hypoadenosinergic state might either be due to a reduction of the extracellular levels of adenosine or alterations in the density of adenosine A 2A receptors (A 2A Rs) or their degree of functional heteromerization with dopamine D 2 receptors (D 2 R). In the present study, we provide preclinical and clinical evidences for this latter mechanism. Two animal models for the study of schizophrenia endophenotypes, namely the phencyclidine (PCP) mouse model and the A 2A R knockout mice, were used to establish correlations between behavioral and molecular studies. In addition, a new AlphaLISA-based method was implemented to detect native A 2A R-D 2 R heteromers in mouse and human brain. First, we observed a reduction of prepulse inhibition in A 2A R knockout mice, similar to that observed in the PCP animal model of sensory gating impairment of schizophrenia, as well as a significant upregulation of striatal D 2 R without changes in A 2A R expression in PCP-treated animals. In addition, PCP-treated animals showed a significant reduction of striatal A 2A R-D 2 R heteromers, as demonstrated by the AlphaLISA-based method. A significant and pronounced reduction of A 2A R-D 2 R heteromers was next demonstrated in post-mortem caudate nucleus from schizophrenic subjects, even though both D 2 R and A 2A R were upregulated. Finally, in PCP-treated animals, sub-chronic administration of haloperidol or clozapine counteracted the reduction of striatal A 2A R-D 2 R heteromers. The degree of A 2A R-D 2 R heteromer formation in schizophrenia might constitute a hallmark of the illness which indeed should be further studied to stablish possible correlations with chronic antipsychotic treatments.