Fig. 29.1
Schematic representation of purine metabolism and different sources of extracellular adenosine. The intracellular ATP could be catabolized into adenosine, which can be further metabolized into inosine and hypoxanthine by intracellular adenosine deaminase (ADA) and purine nucleoside phosphorylase (PNP) respectively. Hypoxanthine can be either salvaged into inosine monophosphate (IMP), or further metabolized to xanthine and uric acid by xanthine oxidase (XO). In addition, adenosine could also be generated intracellularly through the hydrolysis of the S-adenosyl-homo-cysteine (SAH) by an SAH hydrolase. Interestingly, the intracellularly generated adenosine and ATP can be released to the extracellular milieu through an equilibrative nucleoside transporter (ENT) or open-ended systems (i.e., nucleotide-permeable channels, exocytosis, injury or lysis, transport vesicles, lysosomes), respectively. Subsequently, the ATP is dephosphorylated into adenosine by the ectonucleoside triphosphate diphosphohydrolase CD39 and the 5′-nucleotidase CD73, thus constituting the main mechanism behind high extracellular adenosine levels. Finally, extracellular adenosine can be either transported back into the cell through ENTs or transformed into inosine through ecto-adenosine deaminase ADA
Adenosine consists of a purine base (adenine) attached to the 1′ carbon atom of ribose (Fig. 29.1). As mentioned above, this ribonucleoside is mostly produced by the catabolism of ATP, both at the intra- and extracellular levels (Fig. 29.1), although to a lesser extent it can also be generated by S-adenosyl-L-homocysteine (SAH) metabolism (Fig. 29.1). Adenosine, once synthesized, can be either released through Na+-dependent transporters or intracellularly phosphorylated to form AMP by the action of adenosine kinase (Fig. 29.1). In addition, adenosine can react with L-homocysteine to form SAH (Fig. 29.1). Finally, adenosine can be deaminated to form inosine by the action of intra- and extracellular adenosine deaminase (Fig. 29.1).
Adenosine has been historically considered a retaliatory metabolite that “increases oxygen supply and decreases oxygen consumption” [8], thus modulating a large array of physiological processes. Therefore, adenosine participates in the control of respiratory function [9], neural activity [10], platelet aggregation [11], neutrophil function [12], lymphocyte differentiation [13], and vascular tone [14]. Also, adenosine is able to provoke both dilatation of coronary arteries and contraction of kidney blood vessels, thus reducing renal filtration [15]. In addition, it exerts a negative chronotropic and dromotropic effect on the heart [16], as well as mediating the inhibition of neurotransmitter release [17] and lipolysis [18]. Accordingly, adenosine has been postulated as a mediator of metabolic distress, thus having a considerable impact on homeostatic cellular functioning.
Interestingly, within the CNS adenosine has been shown to play a key regulatory role, thus acting as a presynaptic, postsynaptic and/or non-synaptic neuromodulator [19]. Extracellular adenosine levels in the brain range high nM concentration at basal conditions and correlate to the intracellular concentration of adenosine and nucleotides, such as ATP, AMP, and cAMP [20]. Indeed, the intracellular adenosine concentration is related to the rate of breakdown and synthesis of ATP [20]. Thus, adenosine is released as a neuromodulator [21] by the effector cells in response to an increased metabolic demand [22]. Interestingly, it has been postulated that the main source of extracellular adenosine in the striatum comes from intracellular cAMP [23], which is metabolized to AMP by means of phosphodiesterases and then to adenosine by the ecto-nucleotidases (Fig. 29.1). Overall, since cAMP can only be generated by the action of the enzyme adenylyl cyclase, striatal extracellular adenosine would mostly reflect an increased activation of receptors positively linked to adenylyl cyclase.
Adenosine Receptors in the Brain
Early in the 1970s, it was shown that the electrical stimulation of brain slices promoted adenosine release [24]. Interestingly, this stimulated release of endogenous adenosine concomitantly produced cAMP intracellular accumulation, a fact that was blocked by methylxanthine (i.e., caffeine and theophylline) incubation [25]. Moreover, this phenomenon was observed in other tissues (i.e., heart) [26]. Together, these observations constituted the first evidence suggesting that extracellular adenosine exerted its effects through specific plasma membrane receptors. Subsequently, it was demonstrated that the adenosine-mediated antilipolytic effect on fat cells took place with a concomitant reduction in cAMP [27]. Thus, this dual effect of adenosine on cAMP accumulation was further supported when it was confirmed that adenosine could either inhibit or stimulate adenylyl cyclase. Collectively, these observations ended with the first sub-classification of plasma membrane adenosine receptors into Ri and Ra [28], or alternatively, A1 and A2 adenosine receptors [29].
Currently, it is well established that adenosine mediates its actions by activating specific G protein-coupled adenosine receptors (AR), for which four subtypes (A1R, A2AR, A2BR and A3R) have been identified so far. These ARs have a distinguishable pharmacological profile, tissue distribution, and effector coupling [30], and their functioning have been largely studied in the CNS (Table 29.1). ARs belong to the rhodopsin family or class A of G protein-coupled receptors (GPCRs) [52], thus sharing some common molecular signatures. For instance, within their sequence all adenosine receptors contain the widely conserved NPxxY(x)5,6F and the DRY motifs [53, 54]. Thus, adenosine-mediated AR conformational rearrangement determines the binding and activation of specific G proteins (Table 29.1), which are responsible for activation of different intracellular signaling pathways associated with adenosine function (Table 29.1).
Table 29.1
Adenosine receptors
Receptor | Adenosine affinity (nM) | G protein | Transduction mechanismsb | Physiological actions |
|---|---|---|---|---|
A1 | ~70 | Gi/o a Gq/11 Gs | Inhibits: ACa Activates: PLC, AC | |
A2A | ~150 | Gs a Golf G15,16§ | Activates: ACa, PLC Inhibits: Ca2+ channels | |
A2B | ~5,000 | Gs a Gq/11 | Activates: ACa, PLC | |
A3 | ~6,500 | Gi/o a | Inhibits: ACa Activates: PLC |
A1Rs and A2ARs are primarily responsible for the CNS effects of adenosine (Table 29.1) [55]. The most abundant and homogeneously distributed AR within the brain is the A1R, which couples to members of the pertussis toxin-sensitive G proteins (Gi1, Gi2, Gi3 and Go), and whose activation regulates several intracellular effector molecules such us adenylyl cyclase (AC), Ca2+ channels, K+ channels, and phospholipase C (PLC) (Table 29.1) [56]. On the other hand, A2AR is expressed at high levels only in some specific brain regions, for instance striatum, olfactory tubercle, and nucleus accumbens [23, 57]. A2ARs are mainly coupled to Gs/Golf proteins [58], thus activating AC and increasing intracellular cAMP levels (Table 29.1). Interestingly, A2AR may also signal through a G-protein independent pathway eventually associated to mitogen-activated protein kinase (MAPK) signaling cascade activation [59]. Next, the A2BR is positively coupled to AC and PLC through a Gs and Gq protein, respectively (Table 29.1) [2]. A2BR is thought to be fairly ubiquitous in the brain, and the association of this receptor to specific physiological or behavioral responses remains quite scarce, since the A2BR-specific pharmacological tools still are under development [60]. Finally, the A3R has been shown to be coupled to Gi/o proteins, thus inhibiting AC and also stimulating PLC (Table 29.1) [2].
The Adenosine Hypothesis of Schizophrenia
Schizophrenia is a serious mental disorder which comprises a heterogeneous group of syndromes of unknown etiology. It affects up to 1% of the population worldwide, and usually arises at late adolescence and early adulthood (i.e., median age onset is about 23 years in men and 28 years in women) [61]. Importantly, the definition of schizophrenia has evolved through the six editions of the Diagnostic and Statistical Manual of Mental Disorders (DSM) published by the American Psychiatric Association. Thus, for instance, in the DSM-IV version published in 1994, schizophrenia was defined as a mental disorder involving a range of cognitive and emotional dysfunctions that include perception, inferential thinking, language and communication, behavioral monitoring, affect, fluency and productivity of thought and speech, hedonic capacity, volition and drive, and attention. The diagnosis involved the recognition of a constellation of signs and symptoms associated with impaired occupational or social functioning: and no one symptom was pathognomonic of the disorder. In addition, in this version the pathology had very high diagnostic stability, with 80–90% of individuals receiving an initial diagnosis of schizophrenia retaining that diagnosis at 1–10 years [62, 63]. On the other hand, in the current DSM-5 version, five characteristic symptoms for the diagnosis of schizophrenia are established, with the requirement that at least two of these symptoms have to be present for a month [64]. Three changes with respect the previous version have been made, which include the elimination of the special treatment of bizarre delusions and Schneiderian “first-rank” hallucinations, clarification of the definition of negative symptoms, and the addition of a requirement that at least one of the minimum two requisite characteristic symptoms must be delusions, hallucinations, or disorganized speech [64]. It should be noted that the present classification seeks to incorporate the new information about the nature of the disorder accumulated over the past two decades. Thus, the disease is now considered to be characterized by positive, negative, and cognitive symptoms. Positive symptoms reflect the appearance of some phenomena that were not present in the past, and include hallucinations and delusions. On the other hand, negative symptoms, such as anhedonia or apathy, reflect the loss of capacities or characteristics previously possessed. Finally, the cognitive symptoms include alterations in attention, working memory, executive functions, and social cognition.
Numerous theories about the neurotransmission systems affected in schizophrenia have been postulated. Thus, almost all major neurotransmission systems (i.e., dopaminergic, glutamatergic, serotoninergic, GABAergic, and cholinergic) have been involved in schizophrenia, although none of these hypotheses fully explains all the pathological process(es) associated with the disease. Interestingly, one of the most sustained theories is based on a concomitant hyperdopaminergic–hypoglutamatergic phenomenon [65], even though at the beginning they were postulated as separate hypotheses (i.e., the “glutamatergic” and the “dopaminergic” hypothesis) [66, 67]. Indeed, current pharmacotherapy for schizophrenia is based on such a “dopamine” and “glutamate” hypothesis, which focus on a dopamine D2 receptor (D2R) hyperfunction in the striatum, a deficient stimulation of dopamine D1 receptors (D1Rs) in the prefrontal cortex (PFC), and a N-Methyl-D-aspartate (NMDA) receptor hypofunction in the PFC [68]. However, since negative symptoms, cognitive dysfunction, and decrements in psychosocial and vocational functioning are often still persistent upon available pharmacotherapy, the development of a next generation of pharmacologic agents tackling these resilient symptoms is needed [69]. Overall, more research based on non-dopaminergic and non-glutamatergic interventions will be necessary in order to improve the caveats in schizophrenia treatment.
As abovementioned, adenosine plays an important role in the CNS both as a homeostatic neuronal bioenergetic mediator and as a neuromodulator agent. Indeed, an adenosine-mediated modulation of dopaminergic and glutamatergic neurotransmission has been described [70–72]. Thus, adenosine agonists and antagonists produce behavioral effects similar to dopamine antagonists and dopamine agonists respectively [73]. In addition, adenosine tone can also modulate glutamatergic neurotransmission [74, 75]. Therefore, adenosine may play a unique role in integrating glutamatergic and dopaminergic neurotransmission systems, and thus the purinergic hypothesis of schizophrenia has been proposed [76]. Accordingly, early on Lara and co-workers postulated that a dysfunction in the purinergic system (i.e., reduced adenosinergic activity) would account for the imbalance observed between dopaminergic and glutamatergic neurotransmission, a phenomenon that would explain the schizophrenic phenotype [77]. Importantly, and in support of this hypothesis, cognitive impairments and anatomical changes related to psychotic symptoms were recently demonstrated in a mice lacking A2AR [78]. Overall, several elements of experimental evidence supported the adenosine hypothesis of schizophrenia, which will be reviewed here.
Preclinical Models of Schizophrenia: A Role for the Purinergic System?
There is still a considerable lack of knowledge about psychiatric illnesses in general, and schizophrenia in particular. Therefore, preclinical studies based in animal models mimicking some of the schizophrenia-associated symptoms may be useful, even in the case that they do not precisely mirror what exactly occurs in a schizophrenic human patient. Accordingly, preclinical models can be valuable experimental tools to shed light into the mechanisms behind the etiopathology of schizophrenia. As mentioned above, the adenosinergic hypothesis of schizophrenia was proposed to interconnect the schizophrenia-associated dopaminergic hyperfunction and glutamatergic hypofunction. Indeed, some evidence obtained from experimental animal models supported this adenosine contribution to schizophrenia, through the modulation of both dopaminergic and glutamatergic neurotransmission [79]. Hence, we will review here these animal models supporting the glutamatergic hypofunction theory (e.g., the phencyclidine model) and the hyperdopamergic hypothesis (e.g., the amphetamine model) and its relationship with adenosinergic neurotransmission.
The hypoglutamatergic-NMDA receptor hypothesis was formulated in the late 1950s, when it was observed that phencyclidine (PCP) provoked a psychotic-like condition similar to that observed in schizophrenic patients [80]. However, nobody suspected that NMDA receptors were behind this phenomenon until the 1980s, when Lodge and colleagues [81] demonstrated that NMDA receptor blockade was in fact the primary mechanism of PCP-mediated psychotic actions. Indeed, blockade of NMDA receptors promoted both glutamate and dopamine release in the PFC [82], thus disrupting glutamatergic and dopaminergic neurotransmission in this brain region. It has since been postulated that this neurotransmitter unbalance may well be correlated with the cognitive and behavioral perturbations observed in schizophrenia [82–84]. Interestingly, the administration of NMDA receptor antagonists either in the late foetal or in the postnatal period of rats was shown to increase neuronal death by apoptosis [85], a phenomenon that would be linked to adult schizophrenia-like behaviour. Conversely, administration of the same kind of compounds in the adult animal increased the neuronal damage by necrosis with the subsequent gliosis [86], also associated with psychotic-like behaviour. Overall, these experimental observations supported a neurodevelopmental link between NMDA receptor antagonists and schizophrenia. Thus, the hypoglutamatergic-NMDA receptor theory postulates the existence of disturbances in the pre- and perinatal brain development that could provoke clinical manifestations in early adult life [87]. Nevertheless, despite the experimental evidence and some clinical observations, the precise mechanism involving NMDA receptors in schizophrenia is still unknown.
The hypofunction of NMDA receptors in adults, core of the glutamatergic hypothesis of schizophrenia [88], has been traditionally sustained by pharmacological animal models using NMDA receptor antagonists (i.e., PCP, ketamine, and dizocilpine). PCP is a dissociative drug firstly synthesized in 1926 as a surgical anesthetic. Despite its efficacy, the use of this drug was not extended because of its concomitant adverse effects (i.e., hallucinations, delusions, and agitation). Thus, since PCP mimics some schizophrenia symptoms in humans it has been extensively used in animals as a model of this illness. Indeed, in rodents, an acute administration of PCP produced hyperlocomotion [89], social withdrawal [90], and failures both in cognition [91] and in sensorimotor gating [92]. On the other hand, chronic PCP treatment also promoted hyperlocomotion (i.e., a positive symptom) and induced deficits in social behaviour and reduced mobility in the forced swimming test (i.e., negative symptoms). As for the cognitive symptoms, PCP-treated animals displayed sensorimotor gating deficits and cognitive dysfunctions when subjected to learning and memory tests [93]. Interestingly, in humans these PCP-mediated schizophrenic-like symptoms were maintained during several weeks after the chronic treatment [94, 95]. Therefore, the PCP-induced model of schizophrenia seems to partially mimic the pathology, although there also exist some criticisms to this PCP-based animal model. For instance, in animals, differently from the human disease, sensorimotor deficit in the prepulse inhibition test does not last after PCP withdrawal; and with regard to negative symptoms some discrepancies have been reported between clinical features and PCP-treated animals [93].
Importantly, adenosine receptors have been shown to modulate psychostimulant effects in PCP-treated animals. Hence, both A1R and A2AR agonists (i.e., CPA and CGS21680 respectively) were able to counteract PCP-mediated hyperlocomotor activity [96, 97], while A2AR blockade, but not A1R, prompted exacerbation of the motor-stimulant effects of the NMDA antagonist [98]. Indeed, PCP-induced psychomotor activities were enhanced in a KO mouse specifically lacking the striatal neuronal A2AR [99]. However, in a KO mouse lacking the forebrain A2AR, thus with the A2AR deleted in the neurons of striatum as well as cerebral cortex and hippocampus, an opposite effect was observed. Thus, a critical role of A2ARs in extrastriatal neurons was described in providing a major excitatory effect on psychomotor activity [99]. Overall, these results indicate that A2ARs in striatal and extrastriatal neurons exert an opposing modulation of psychostimulant (i.e., PCP-mediated) effects.
Similarly, the dopaminergic hypothesis of schizophrenia had several important theoretical changes throughout its history. Thus, while at the beginning it was based on a generalized hyperdopaminergic brain function, it quickly evolved into a combined subcortical hyperdopaminergic-prefrontal hypodopaminergic dysfunction. However, Howes proposed an updated third version based on multiple changes of different neurotransmitters and neural systems, which together with other biological or environment influences would underlie the cognitive dysfunction and negative symptoms of schizophrenia. In Howes’ words, rather than being a hypothesis of schizophrenia this new view is more accurately a “dopamine hypothesis of psychosis-in-schizophrenia”. This hypothesis explains several environmental and genetics risks of schizophrenia, and proposes that these interact to funnel through one final common pathway of presynaptic striatal hyperdopaminergia [100].
The hyperdopaminergic status of schizophrenia has been largely studied by means of pharmacological animal models. Thus, the administration of drugs (i.e., amphetamine) increasing the brain dopamine content is a classical experimental approach to mainly study schizophrenia-associated positive symptoms. Amphetamine, first discovered in 1887 [101], is currently used as an attention deficit (i.e., ADHD) and narcolepsy treatment [102]. It is a drug that acts as a strong CNS stimulant by increasing dopamine concentration in the synaptic cleft and thus raising the response in the post-synaptic neuron. Apart from the well-known positive symptoms, its administration can also provoke long-term cognitive impairments [103, 104]. Overall, while several investigations have demonstrated that amphetamine treatment could induce some behavioral, molecular, cellular, and neurochemical changes, which were behind the striatal dopaminergic system [105–108], studies reporting amphetamine-mediated negative symptoms are rare.
Dopamine receptors on striatonigral and striatopallidal neurons (D1R and D2R respectively) play a pivotal role in the control of motor responses [109]; thus, the efficacy of many antipsychotic drugs correlates well with their ability to block D2Rs [110]. Since A2ARs antagonistically interact with D2Rs [111, 112], adenosine is expected to exert a regulatory influence on psychomotor behaviour, and indeed a role for A2AR regulating amphetamine-induced psychomotor behaviour has been described [113]. Thus, A2AR activation restored responsiveness to amphetamine in adenosine-deficient mice [113]. Overall, the abovementioned preclinical data supported the involvement of adenosine in schizophrenia and the potential use of adenosine receptors as drug targets for this disease.
Clinical Evidence Supporting the Adenosine Hypothesis
Several lines of investigation support the notion that the adenosinergic system may be altered in schizophrenia. The first remarkable piece of information pointing to this consists of the discovery that A2AR expression was found to be altered in necropsies from schizophrenic individuals. Thus, A2AR binding was increased in the striatum of postmortem brains of chronic schizophrenics [114, 115]. Similarly, an increased expression of A2AR on perivascular astrocytes in the hippocampus of patients with schizophrenia has been described [116]. On the contrary, a reduced expression of A2AR in human postmortem putamens of patients suffering schizophrenia has been reported, thus proposing that there may be a subgroup of schizophrenic patients with reduced striatal A2AR levels accompanied by an altered motor phenotype [117]. Indeed, since the adenosinergic tone was shown to be altered in schizophrenia, A2AR up- and down-regulation in a brain region-dependent manner may correspond to adaptive physiological conditions that in turn would be associated to a concomitant hyperdopaminergic state [77]. Interestingly, the genetic linkage of adenosine receptors to schizophrenia has been evaluated. For instance, while an A2AR genetic variant (i.e., 1976 T > C) was not shown to confer susceptibility to schizophrenia [118] an A1R gene polymorphism was associated with pathophysiological mechanisms underlying the schizophrenia, thus becoming a potential useful biomarker of schizophrenia [119]. In addition, the most frequent functional polymorphism of adenosine deaminase (22G → A, ADA1*2), which is characterized by a reduced enzymatic activity and thus higher adenosine levels, is less frequent among schizophrenic patients [120]. Overall, these results support the hypothesis of lower adenosinergic activity in schizophrenia.
Based on the previous data, it seems feasible to think that the use of pro-adenosinergic drugs may be beneficial for the treatment of the pathology. However, this pharmacological proposal is still premature, although some data exist concerning this suggestion. Indeed, raising the endogenous pool of purines with allopurinol has been shown to produce some promising results as add-on therapy for schizophrenia [121, 122]. Allopurinol, a well-known hypouricemic drug that inhibits xanthine oxidase (Fig. 29.1 ) was used as an add-on drug in the treatment of poorly responsive schizophrenic patients [121]. Interestingly, in this short controlled trial (i.e., 23 patients treated with haloperidol 15 mg/day plus allopurinol 300 mg/day and 23 patients with haloperidol 15 mg/day plus placebo) it was observed that the combination of haloperidol and allopurinol showed a significant superiority over haloperidol alone in the treatment of positive symptoms and general psychopathology symptoms, as well as Positive and Negative Syndrome Scale (PANSS) total scores [121]. In a similar study, a double-blind, placebo-controlled, crossover clinical trial of add-on allopurinol (300 mg/day) for poorly responsive schizophrenia or schizoaffective disorder (DSM-IV criteria), was conducted in 22 patients [122]. In this case, allopurinol was an effective and well-tolerated adjuvant treatment, especially for refractory positive symptoms [122]. Also, allopurinol showed effectiveness as adjunctive medication in schizophrenia outpatients (N = 59) with persistent symptoms despite adequate pharmacotherapy [123]. And more recently in a case report, allopurinol prompted a rapid decrease in psychotic symptoms in a patient with schizophrenia [124]. Thus, within 2 weeks of allopurinol adjuvant therapy, the patient showed significant improvement with respect to his positive and negative symptoms of schizophrenia (PANSS scores went from 88 to 41 over a period of 2 weeks) [124]. Although some clinical controversy has been established around allopurinol [125], adenosine modulator adjuvant therapy was shown to be more beneficial in overall psychopathology (especially positive symptoms) in schizophrenia and in treating mania episodes of bipolar disorder when compared to placebo [126]. Overall, these clinical studies suggest that allopurinol might be an effective adjuvant drug in the management of patients with chronic schizophrenia who are poorly responsive to current treatments. However, larger, randomized clinical trials are needed before a broad clinical application of allopurinol is recommended as routinely used adjuvant therapy to antipsychotics [127].
Another piece of evidence supporting the link between the adenosinergic system and schizophrenia consists of the fact that the adenosine transport inhibitor dipyridamole was found to be beneficial in patients with schizophrenia [128]. Thus, raising extracellular adenosine levels with dipyridamole not only improved haloperidol-mediated amelioration of positive and general psychopathology symptoms, as well as PANSS total scores [128], but it also showed effectiveness when combined with lithium in the treatment of acute bipolar mania [129]. Overall, all the above-mentioned clinical data support the adenosine hypothesis of schizophrenia and highlight the potential pharmacological interest of combining antipsychotic drugs with purinergic-based compounds (i.e., allopurinol and dipyridamole) to tackle resilient schizophrenia symptoms.
Related posts:
Life Project
Dialogues Between Philosophy and Psychiatry: The Case of Dissociative Identity Disorder
Psychoneuroendocrinological and Cognitive Interactions in the Interface Between Chronic Stress and Depression
Eye Movements: Parameters, Mechanisms, and Active Vision
Metabolic Association Between the Gut–Brain Axis in Autism Spectrum Disorders
Liaison Psychiatry: Playing “Hide and Seek” with Delirium
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
