**Background:** The adenosine hypothesis of schizophrenia, formulated in the early 2000s, postulates that brain adenosine deficiency can cause schizophrenia symptoms and that strengthening adenosinergic signaling should be therapeutic. Adenosine acts through four G-protein-coupled receptors (ADORA1, ADORA2A, ADORA2B, ADORA3), with ADORA1 and ADORA2A being the primary targets under physiological conditions. ADORA1 inhibits dopamine and glutamate release and suppresses neuroplasticity, while ADORA2A facilitates neurotransmitter release and is involved in long-term potentiation. The functional antagonism between adenosine and dopamine at ADORA2A/D2R heteromers has led to suggestions that ADORA2A stimulation may confer antipsychotic potential. This review aims to evaluate clinical evidence for adenosine hypofunction in schizophrenia and explore adenosine-dependent immune and epigenetic processes from an etiological perspective.
**Methods:** This is a narrative review synthesizing clinical and preclinical evidence related to the adenosine hypothesis of schizophrenia. The authors review post-mortem studies, genetic analyses, enzymatic activity measurements, PET imaging studies, clinical trials of adenosine-augmenting drugs, and animal models. They examine multiple components of the adenosine system: adenosine kinase (ADK), adenosine deaminase (ADA), ectonucleotidases (EctoNTs), equilibrative nucleoside transporters (ENTs), and adenosine receptors (ADORA1, ADORA2A). The review also explores adenosine's role in mediating environmental risk factors including maternal immune activation, perinatal hypoxia, and caffeine exposure during pregnancy.
**Key Results:** (1) ADK expression: Multiple post-mortem studies found no significant changes in ADK mRNA or protein levels in the dorsolateral prefrontal cortex (DLPFC) or anterior cingulate cortex of schizophrenia patients. Cell-type specific measurements and in silico analyses confirmed null findings. (2) ADA activity: Increased ADA gene expression was detected in pyramidal neurons of the DLPFC. Serum ADA activity was elevated in drug-naïve first-episode patients and in medicated patients, with clozapine-treated patients showing higher activity than haloperidol-treated patients. A rare ADA polymorphism with lower enzymatic activity was less frequent in a Brazilian cohort of schizophrenia patients. (3) EctoNTs: Two post-mortem studies reported reduced EctoNT activity in the striatum and downregulation of EctoNT mRNA expression in astrocytes in the DLPFC. However, preclinical models showed that genetic deletion of CD73 (a key EctoNT) blunted amphetamine sensitization and improved working memory, contradicting the expectation that reduced EctoNTs contribute to symptom production. (4) ENTs: Reduced ENT1 expression was found in the superior temporal gyrus and in pyramidal neurons of the DLPFC. (5) Adenosine receptors: ADORA2A expression increases in hippocampus and striatum were reported but appear attributable to antipsychotic medication. A PET imaging study with [11C]SCH442416 failed to differentiate 12 medicated chronic schizophrenia patients from 13 healthy controls. One study reported reduced ADORA2A-D2R heterodimer formation in the caudate nucleus. (6) Clinical trials: Three adenosine-augmenting drugs (allopurinol, dipyridamole, propentofylline) showed limited efficacy as adjuncts to antipsychotics, with meta-analysis showing small effect sizes largely limited to positive symptoms. A Finnish longitudinal study (N=61,889) found that allopurinol and dipyridamole add-on reduced re-hospitalization risk between 1996 and 2017. (7) Caffeine: Higher coffee consumption is reported in schizophrenia patients. Acute caffeine improved semantic fluency, processing speed, working memory, and visual memory in male patients only. (8) RNA editing: ADAR and ADARB1 expression was upregulated in the anterior cingulate cortex and DLPFC in post-mortem samples. In a mouse model of prenatal maternal immune activation (Poly(I:C) at gestational day 9), A-to-I RNA editing increased in fetal brains 24 hours post-challenge, accompanied by increased expression of the interferon-inducible 150-kDa ADAR isoform. (9) Caffeine during pregnancy: A Danish national birth cohort study (N=47,491) found maternal caffeine consumption during pregnancy associated with increased risk of psychiatric disorders in offspring. Mouse studies showed prenatal caffeine exposure induced working memory deficits and potentiated locomotor response to cocaine in adult offspring.
**Clinical Implications:** The evidence for adenosine hypofunction in schizophrenia is mixed and regionally complex. ADK expression appears largely unaffected, while ADA activity is elevated and EctoNT/ENT expression is reduced. Adenosine receptor changes are confounded by medication effects. Current adenosine augmentation therapy shows limited efficacy, primarily for positive symptoms, with no benefits for negative or cognitive symptoms. The authors suggest shifting focus from symptom suppression to etiological intervention, particularly regarding adenosine's role in mediating early-life environmental risks. Adenosine augmentation during pregnancy may potentially protect against neurodevelopmental risks from maternal infection, though this must be balanced against risks of excessive immunosuppression. Targeting ADAR-mediated RNA editing and DNA methylation through adenosine regulation represents novel therapeutic avenues. The possibility that ADORA2A antagonists (rather than agonists) may benefit cognitive symptoms, consistent with heavy caffeine consumption by patients as potential self-medication, challenges the core adenosine hypofunction hypothesis and warrants further investigation.