Pathological overproduction: the bad side of adenosine

Abstract

Adenosine is an endogenous ubiquitous purine nucleoside, which is increased by hypoxia, ischaemia and tissue damage and mediates a number of physiopathological effects by interacting with four GPCRs, identified as A₁, A_2A, A_2B and A₃. Physiological and acutely increased adenosine is mostly associated with beneficial effects that include vasodilatation and a decrease in inflammation. In contrast, chronic overproduction of adenosine occurs in important pathological states, where long‐lasting increases in the nucleoside levels are responsible for the bad side of adenosine associated with chronic inflammation, fibrosis and organ damage. In this review, we describe and critically discuss the pathological overproduction of adenosine and analyse when, where and how adenosine exerts its detrimental effects throughout the body.

Abbreviations

AD

Alzheimer's disease

ADK

adenosine kinase

Aβ

amyloid‐β

BBB

blood–brain barrier

CD39

apyrase

CD73

5′‐nucleotidase

COPD

chronic obstructive pulmonary disease

EAE

experimental autoimmune encephalomyelitis

HD

Huntington's disease

HIF

hypoxia‐inducible factor

IBD

inflammatory bowel diseases

KO

knockout

L‐DOPA

levodopa

MDSC

myeloid‐derived suppressor cells

MS

multiple sclerosis

NTs

nucleoside transporters

PD

Parkinson's disease

RLS

restless legs syndrome

αSyn

α‐synuclein

Tables of Links

TARGETS
GPCRs a	Enzymes b	Transporters c
A1 receptor	5’‐nucleotidase	Glutamate transporter 1
A2A receptor	Adenosine deaminase	SLC 28 and 29 nucleoside transporters
A2B receptor	Adenosine kinase
A3 receptor	Apyrase (CD39)
D1 receptor	Caspase 1
D2 receptor	ERK1
	ERK2
	JNK
	P38
	PLC
	PKA

LIGANDS
Adenosine
ADP
ATP
cAMP
CXCL8
Dopamine
IL‐2
IL‐6
IL‐13
IL‐17A
VEGF‐A

These Tables list key protein targets and ligands in this article which are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Southan et al., 2016), and are permanently archived in the Concise Guide to PHARMACOLOGY 2015/16 (^a,b,cAlexander et al., 2015a,b,c).

Introduction

Adenosine is an endogenous ubiquitous autacoid regulating the function of every tissue and organ in the body (Borea et al., 2016). It is produced both intra‐ and extracellularly following the activity of specific enzymes, and its extracellular concentration is tightly regulated within a range of 30–200 nM (Fredholm et al., 2011). In particular, the main extracellular source of adenosine is the degradation of ATP and AMP by the ectoenzymes apyrase (CD39) and 5′‐nucleotidase (CD73), respectively, whilst for the intracellular level, its generation depends on AMP hydrolysis by an intracellular CD73 or S‐adenosylhomocysteine hydrolase (Zimmermann, 2000; Eltzschig, 2009). Adenosine concentration is maintained in equilibrium through equilibrative solute carrier family 29 (SLC 29) or concentrative solute carrier family 28 (SLC 28) nucleoside transporters (NTs) present in the cell membrane or through its phosphorylation operated by an intracellular adenosine kinase (ADK) (Alexander et al., 2015c). In addition, adenosine signalling can be arrested following adenosine deaminase activity, which is present both inside and outside the cell, transforming adenosine to inosine. Extracellular adenosine increases up to μM levels in hypoxia/ischaemia due to the up‐ and down‐regulation of CD73 and ADK, respectively; this increase in adenosine is characteristic of several important pathological conditions as it is involved in the regulation of energy supply/demand (Fredholm, 2014). Due to its rapid metabolism, adenosine behaves like an autocrine/paracrine rather than a systemic mediator (Figure 1).

Figure 1.

Schematic representation of the adenosinergic system from adenosine formation and release to signalling and removal from the extracellular space. Alterations occurring during hypoxia are shown. 5′NT, 5′‐nucleotidase; ADA, adenosine deaminase; SAH, S‐adenosylhomocysteine.

Adenosine interacts with four adenosine receptors named A₁, A_2A, A_2B and A₃ belonging to the family of GPCRs (Borea et al., 2015). All of them affect cAMP levels, A₁ and A₃ receptors being coupled to inhibitory Gi and A_2A and A_2B receptors to stimulatory Gs proteins respectively. Furthermore, A₃ and A_2B receptors are also coupled to Gq proteins, leading to the activation of PLC and an increase in calcium levels. Upon activation of A₁ and A₃ receptors, K_ATP channels are opened. Furthermore, all adenosine receptors stimulate MAPKs, triggering a variety of intracellular signalling pathways relevant for a number of cell functions including growth and proliferation, apoptosis, necrosis and inflammation (Schulte and Fredholm, 2003).

The wide distribution of adenosine receptors in almost all cells of the organism makes them an attracting target for drug development in several diseases, even though this ubiquity makes the development of selective adenosine receptor agonists and antagonists difficult to realize without the occurrence of side effects (Chen et al., 2013). Recently, we have published a review about the role of adenosine as a guardian angel against cellular damage, highlighting where, when and how adenosine exerts its protective mission under both physiological and pathophysiological conditions, in response to stress, both in the brain and in the periphery (Borea et al., 2016). However, with regard to the ‘guardian angel’ metaphor, we should not forget that there are instances in which a chronic overproduction of adenosine is pathological, particularly in Parkinson's disease (PD), fibrosis, hepatic steatosis, colitis, asthma, diabetes and cancer.

In this review, we focus on the pathological overproduction of adenosine describing when, where and how it exerts its detrimental effects throughout the body.

Pathological role of adenosine in cerebral ischaemia

In the CNS, both neurons and glial cells produce adenosine and express adenosine receptors that affect several homeostatic functions ranging from sleep to arousal, learning, memory and cerebral blood circulation. The principal adenosine receptor subtypes located in the CNS are A₁ and A_2A receptors, which are crucial in the regulation of neurotransmitter release, neuronal excitability, synaptic plasticity, vasodilatation and neuroinflammation (Sebastiao and Ribeiro, 2009). In general, the role of A₁ receptors emerges as neuroprotective being an inhibitor of excitatory transmission, while that of the A_2A receptor is detrimental, even though evidence has been obtained showing that its activation has beneficial effects (Melani et al., 2014; Pedata et al., 2016). The molecular bases of the adverse effects of A_2A receptor stimulation are attributed to its modulation of neuronal glutamate release, potentiation of NMDA‐mediated effects, central inflammatory processes, glial reactivity, blood–brain barrier (BBB) permeability and infiltration of peripheral immune cells, which are associated with excitotoxicity and the pathogenesis of several brain diseases (Yu et al., 2004, 2008; Chen et al., 2008a,b; Azdad et al., 2009; Carta et al., 2009; Gui et al., 2009; Melani et al., 2009; Dai et al., 2010). Serendipitous pioneering studies about the involvement of A_2A receptors in ischaemia found that its antagonism protected the brain against ischaemic insults (Gao and Phillis, 1994; Lubitz et al., 1995; Phillis, 1995; Monopoli et al., 1998). Later, this finding was confirmed in A_2A receptor knockout (KO) mice and/or following pharmacological inhibition of the receptor, where a reduction in infarcted area was observed following ischaemic damage (Chen et al., 1999; 2007; Chen and Pedata, 2008). It has been demonstrated that after ischaemia, excitotoxicity is the first phenomenon occurring in the brain in the first 4 h, when the A_2A receptor is responsible for the increase in glutamate levels, through both the release of glutamate from glutamatergic terminals and the inhibition of the glutamate‐1 transporter (GLT‐1) in astrocytes. Indeed, it has been shown that A_2A receptor blockade protects against excitotoxicity (Matos et al., 2012b, 2013). Furthermore, A_2A receptor activation reduces the affinity of agonists for A₁ receptors in A₁–A_2A receptor heteromers, which exert a presynaptic control of striatal glutamate release, resulting in the fine‐tuning of the modulatory effect on striatal glutamatergic neurotransmission (Ciruela et al., 2006).

At the intracellular level, the inhibition of p38 MAPK in microglia and JNK in oligodendrocytes may be responsible for the protective effect mediated by A_2A receptor antagonists in ischaemia (Melani et al., 2006, 2009). However, it is also known that later, hours and days after insult, there is a massive infiltration of blood cells and neuroinflammation that may be inhibited by activation of A_2A receptors located on blood cells. Accordingly, the chronic administration of an A_2A receptor antagonist for 7 days after transient focal ischaemia failed to protect the brain (Melani et al., 2015), suggesting that A_2A receptor antagonists should be administered for a limited time window of about 4 h after stroke (Pedata et al., 2016). However, it is important to mention that chronic enhanced levels of adenosine may lead to seizures, through A_2A receptor activation and neurotrophins, suggesting that in this context A_2A receptor antagonists could be useful clinically (Sandau et al., 2016). Interestingly, recent evidence adds a new pathological aspect to adenosine overproduction related to the development of pain behaviour. This effect seems to be due to the activation A_2B receptors on myeloid cells, which are able to transactivate nociceptors of sensory neurons, and induce hypersensitive neurons and chronic pain through a pathway involving IL‐6. Hence, the A_2B receptor may play a major role in chronic pain by promoting immune–neuronal interactions (Hu et al., 2016) (Figure 2).

Figure 2.

Deleterious effects of adenosine mediated through A_2A and A_2B receptors expressed in different areas of the CNS. DRG, dorsal root ganglion.

Pathological role of adenosine in neurodegenerative diseases

There is a well‐established allosteric relationship between striatal A_2A receptors and dopamine D₂ receptors, supported by evidence that A₂ receptor activation is responsible for the decreased affinity of agonist binding to D₂ receptors (Ferrè et al., 1991). This represented the first proof of concept for the therapeutic utility of A_2A receptor antagonists in PD (Armentero et al., 2011; Preti et al., 2015; Ferré et al., 2016) (Figure 2). Interestingly, the well‐known epidemiological consideration that caffeine protects against PD has been confirmed in an animal model of PD and attributed to A_2A receptor antagonism (Ross et al., 2000; Ascherio et al., 2001, 2004; Chen et al., 2001; Yu et al., 2008; Xu et al., 2016). A_2A receptors and D₂ receptor heteroceptor complexes have been detected in cellular models and in the striatum (Hillion et al., 2002; Canals et al., 2003; Azdad et al., 2009; Borroto‐Escuela et al., 2010, 2011, 2013; Varani et al., 2010; Trifilieff et al., 2011; Bonaventura et al., 2015). As the receptors show antagonistic interactions at both membrane and intracellular signalling levels, dopamine depletion may be responsible for an overactivation of A_2A receptors and the consequent symptoms of PD (Peterson et al., 2012). Therefore, antagonists of A_2A receptors are beneficial for the improvement in motor function in different animal models of PD, as they attenuate the inhibition exerted via A_2A receptors on the effects of D₂ receptors in the gabaergic striato‐pallidal neurons (Fuxe et al., 2015). Recently, pharmacological data have suggested a new model that indicates A_2A receptors and D₂ receptor heteromers form heterotetramers, which consist of A_2A receptors and D₂ receptor homodimers. This model supports the evidence that, at high concentrations, A_2A receptor antagonists exert the same effects as A_2A receptor agonists, thus reducing D₂ receptor‐mediated activity in neurons; these findings could be important from a clinical point of view, when the use of A_2A receptor antagonists is being considered for therapy (Bonaventura et al., 2015). Even though the development of A_2A receptor antagonists as new drugs for PD treatment has encountered several obstacles in the clinic, istradefylline has recently been registered in Japan, as it has been found to decrease the “off time” when used in combination with levodopa (L‐DOPA). However, it has not receive approval by the American Food and Drug Administration, due to the lack of significant improvement in the phenomenon of “wearing off” in comparison with L‐DOPA. Other potential new drugs like preladenant and vipadenant have failed in phase III clinical studies for PD (Navarro et al., 2016; Oertel and Schulz 2016). The reasons for which these drugs failed were the lack of significant benefits of preladenant combined with L‐DOPA when compared to L‐DOPA alone, while vipadenant was found to have toxic effects. It has been hypothesized that administration of this class of compounds should start as early as possible to avoid a structural change in the A_2A receptor–D₂ receptor heteroceptor complexes that may induce dyskinesias and the onset of “off time” caused by L‐DOPA (Fuxe et al., 2015). Another pharmacological approach for PD treatment is suggested from the results obtained with bivalent drugs acting at A_2A receptor–D₂ receptor complex (Soriano et al., 2009), with integrated dual acting ligands having an improved efficacy in preliminary BBB permeability tests (Jörg et al., 2015). Recently, a novel protective effect of A_2A receptor antagonists has been attributed to the modulation of α‐synuclein (αSyn) effects, as α‐synuclein‐induced damage to striatal neurons was clearly reduced in A_2A receptor KO mice and A_2A receptor antagonists prevented the neurotoxicity related to α‐synuclein aggregation. These results provide additional evidence in support of antagonists as effective drugs for the treatment of PD and related synucleinopathies (Kachroo and Schwarzschild, 2012; Ferreira et al., 2015). The use of A_2A receptor antagonists has been found useful in other pathologies involving neuronal dysfunction including Huntington's (HD) and Alzheimer's (AD) diseases as well as major depression and schizophrenia, epilepsy, acute and chronic stress, restless legs syndrome (RLS) and memory fear (Cunha, 2005, 2016; Canas et al., 2009; Batalha et al., 2013; Krügel, 2016; Laurent et al., 2016; Quiroz et al., 2016; Simões et al., 2016) (Figure 2). Different studies have reported an up‐regulation of A_2A receptors in animal models of HD (Varani et al., 2001; Tarditi et al., 2006), and recently, it has been shown that A_2A receptor antagonists block working memory deficits at early stages of HD models (Li et al., 2015). Hyperactivation of both D₁ and A_2A receptors has been found in HD striatum where D₁ plus A_2A receptor antagonists reduced PKA activity, which is involved in hippocampal‐dependent cognitive dysfunction in HD, further supporting a therapy based on A_2A receptor blockade. Nevertheless, evidence in favour of A_2A receptor activation as a strategy for HD treatment has also been reported (Gomes et al., 2011; Martire et al., 2013; Tyebji et al., 2015). In addition, a positive action of A_2A receptors in the early phases of neurodegeneration has been attributed to the presynaptic facilitation of GABA transmission exerted by A_2A receptor agonists, involving brain‐derived neurotrophic factor–tropomyosin receptor kinase B in the hippocampus (Colino‐Oliveira et al., 2016).

A change in the pattern of adenosine receptors with a decrease in A₁ and an increase in A_2A receptors has been also detected in AD. As a consequence A_2A receptor antagonists have been found to prevent neuronal death and the decrease in astrocytic glutamate uptake, caused by an amyloid‐β (Aβ) fragment, that may be responsible for excitotoxicity in AD (Dall'Igna et al., 2003; Canas et al., 2009; Matos et al., 2012a). Importantly, it has been shown that humans with AD have increased levels of A_2A receptors in astrocytes. Also, in young and aging mice, their genetic removal increased long‐term memory and increased the levels of an immediate‐early gene necessary for memory (Orr et al., 2015). Furthermore, deletion of A_2A receptors has been shown to have a protective effect from spatial memory and hippocampal long‐term depression caused by Tau pathology (Laurent et al., 2016; Müller et al., 2016). In addition, A_2A receptors are up‐regulated in CA3 synapses at early stages of AD and their antagonism reversed the block of long‐term synaptic potentiation in a mouse model of AD amyloidosis (Viana da Silva et al., 2016). Caffeine has been found to have a protective effect against cognitive impairment in both human and animal studies (Maia and de Mendonça, 2002; Ritchie et al., 2007; Smith, 2009; Santos et al., 2010). Furthermore, caffeine reduced plasma and brain Aβ levels in an animal model of AD and prevented memory deficits caused by Aβ administration (Dall'Igna et al., 2007; Cao et al., 2009). Interestingly, the initial findings of a case–control study were the first to demonstrate that caffeine/coffee consumption is associated with a reduced risk, or delayed onset, of dementia (Cao et al., 2012). Therefore, it seems that caffeine, the most popular and widely used drug in the world, by antagonizing the effects of adenosine, retains a big potential to counteract different neurodegenerative diseases (Woods et al., 2016).

Due to their modulation of glutamatergic and monoaminergic transmission in the striatum, A_2A receptors may also be involved in depression (Krügel, 2016). In behavioural animal models, A_2A receptor antagonists produced antidepressant effects (El Yacoubi et al., 2003; Hodgson et al., 2009). However, istradefylline was found to exert them independently from monoaminergic transmission (Yamada et al., 2014a,b). Furthermore, caffeine or selective A_2A receptor antagonists reversed the performance deficits in reserpine‐treated rats and prevented mood and memory dysfunctions induced by chronic stress (Batalha et al., 2013; Kaster et al., 2015; Minor and Hanff, 2015). In addition, in a broad‐based model of depression mediated by an increase in hippocampal A_2A receptors, which control synaptic glutamatergic function, caffeine was able to prevent memory, but not mood, deficits (Machado et al., 2016).

Overall, based on the progress obtained in clarifying the molecular mechanisms underlying CNS diseases, adenosine reveals its bad side essentially through the activation of A_2A receptors, which may be considered as targets of therapeutic strategies and may represent a promising future in the battle against a wide spectrum of unmet central diseases.

Pathological role of adenosine in autoimmune diseases

Immune cells express all adenosine receptors, which means adenosine can affect inflammatory and immune events (Cekic and Linden, 2016). In particular, it was found that adenosine modulates immune function by exerting inhibitory effects on neutrophils, lymphocytes, monocytes/macrophages and dendritic cells (Mills et al., 2012). Therefore, a possible area of investigation for novel therapeutic strategies through adenosine receptor modulation is the study of adenosine's role in the immunopathogenesis of multiple sclerosis (MS).

Interestingly, it has been observed that mice with CD73 knocked‐down, the critical enzyme for generating extracellular adenosine in various cells including T cells, are protected against experimental autoimmune encephalomyelitis (EAE, an animal model for MS) and lack the CNS lymphocyte infiltrates associated with disease progression (Mills et al., 2008).

Moreover, caffeine decreases the incidence of EAE and attenuates it at the behavioural, histological and neurochemical levels in mice and rats (Tsutsui et al., 2004; Chen et al., 2010). Furthermore, it shifted the status of the cytokine levels from Th1, pro‐inflammatory to Th2, anti‐inflammatory in EAE, then ameliorated inflammatory injury in the brain and spinal cord, suggesting it has a neuroprotective effect via its role as an immunomodulator (Chen et al., 2010).

Surprisingly, even though the A_2A receptor is recognized as a major mediator of anti‐inflammatory responses, it has been reported that the pharmacological blockade of this subtype attenuates EAE pathology in CD73 KO mice (Mills et al., 2008). Furthermore, A_2A receptor blockade prevented oligodendrocyte damage through a reduction of JNK/MAPK signalling in these cells (Melani et al., 2009). However, to explain the beneficial effects of A_2A receptor antagonists in EAE, it has been suggested that although the A_2A receptor present on lymphocytes is anti‐inflammatory, that expressed in the CNS plays a pro‐inflammatory role and appears to be essential for EAE development (Mills et al., 2012). However, an A_2A receptor agonist worsened experimental autoimmune neuritis through an inhibitory effect of T‐cell proliferation and IL‐2 secretion (Zhang et al., 2016). The role of A_2B receptors in the pathogenesis of MS has also been investigated by knocking out this subtype or blocking it with selective antagonists. A_2B receptors have been found to be up‐regulated in peripheral blood leukocytes of MS patients and the peripheral lymphoid tissues of EAE mice. Furthermore, pharmacological blockade of the A_2B receptor ameliorated the clinical symptoms of EAE and reduced immune damage in the CNS. Importantly in A_2B receptor KO mice, EAE was less severe. The mechanism involved was suggested to be the inhibition of Th17 cell differentiation by the blockade of IL‐6 production from dendritic cells, via the PLCβ–PKC and p38 MAPK pathways (Wei et al., 2013). Accordingly, A_2B receptor stimulation in dendritic cells greatly increased the development of experimental autoimmune uveitis (EAU) by enhancing the Th17 autoimmune responses (Chen et al., 2015). In conclusion, adenosine exerts a complex role, acting at A_2A and A_2B receptors to increase inflammation, EAE and EAU.

Pathological role of adenosine in inflammatory conditions

Inflammation is a hallmark of important diseases such as cardiovascular, metabolic, intestinal and pulmonary disorders (Hotamisligil, 2006; Libby et al., 2011; Barnes, 2016; Karin and Clevers, 2016). As for the detrimental signalling of adenosine in inflammatory conditions, we have to focus on A_2B receptors as mediators of the effects of autacoids because this subtype is activated by μM concentrations of adenosine, occurring in inflammation, hypoxia and cell injury (Fredholm, 2007; Bartels et al., 2013). In particular, the expression of A_2B receptors is increased in hypoxia in response to hypoxia‐inducible factors (HIFs) (Eckle et al., 2014). Indeed, it has been demonstrated that A_2B receptors have a role in the inflammatory environment typically present in asthma, chronic obstructive pulmonary disease (COPD), kidney pathologies and inflammatory bowel diseases (IBD) (Kolachala et al., 2005; Haskó et al., 2008). Likewise, A₃ receptors expressed in the human lung, mast cells and in macrophages enriched with lipids, known as foam cells (FC), play a role in the regulation of the inflammation present in pulmonary and atherosclerotic diseases (Hua et al., 2008; Gessi et al., 2010; Borea et al., 2015).

The A_2B receptor is present in the intestinal epithelium where it exerts a controversial role in inflammation (Kolachala et al., 2005). Indeed, different data have reported that A_2B receptors are able to protect from inflammation development in the intestine (Schingnitz et al., 2010; Hart et al., 2011; Aherne et al., 2015). However, the expression of A_2B receptors in non‐immune cells plays an important role in the occurrence of murine colitis. Indeed, colitis is attenuated in A_2B receptor KO mice and after of pharmacological blockade of A_2B receptors, indicating that A_2B receptor antagonism may be an effective treatment for acute inflammatory intestinal diseases, such as IBD (Kolachala et al., 2008a,b; Ingersoll et al., 2012).

Adenosine is responsible for the production of VEGF, IL‐6 and IL‐8 through interaction with all four adenosine receptors in inflamed tissues, in which the levels of these autacoids are increased (Cekic and Linden, 2016; Koszałka et al., 2016). However, the A_2B and A₃ receptor subtypes are mainly involved in the modulatory effects of adenosine on wound healing processes, for example, angiogenesis and fibrosis (Zhou et al., 2011) (Figure 3). The A_2B receptor plays a significant role in the chronic phase of wound healing after tissue injury and increases the adverse signalling responsible for the continuous tissue remodelling and fibrosis occurring in chronic inflammatory conditions (Feoktistov et al., 2009). Pulmonary fibrosis is a harmful lung disease with limited therapeutic options. Adenosine has a role as a pro‐inflammatory messenger in different chronic pulmonary inflammatory conditions, such as asthma and COPD (Cekic and Linden, 2016). Very recently, it has been reported that increased extracellular adenosine concentrations, by inducing the production of IL‐6 and IL‐17, are associated with the progression of experimental pulmonary fibrosis (Luo et al., 2016). As the A_2B receptor promotes Th17 differentiation (Wilson et al., 2011), it has been hypothesized that adenosine may induce IL‐17 expression through its A_2B receptor subtype in chronic lung injury, thus contributing to lung fibrosis. Furthermore, adenosine promotes the differentiation of alternatively activated macrophages, a macrophage subtype that has been shown to contribute to pulmonary fibrosis, stimulated by Th2 cytokines (IL‐4 and IL‐13) (Karmouty‐Quintana et al., 2015). In addition, it has been reported that A_2B receptors are up‐regulated in lung tissue from idiopathic pulmonary fibrosis patients (Zhou et al., 2010) and that the pharmacological blockade of A_2B receptors in mice (Karmouty‐Quintana et al., 2012, 2013b; Sun et al., 2006) or genetic removal of this subtype (Zhou et al., 2011) are associated with decreased fibrosis. Therefore, it is important to verify and understand the association between extracellular adenosine levels and the progression of pulmonary fibrosis to develop new drugs for this pathology based on the structures of adenosine ligands binding to A_2B receptors. As for the role of adenosine in asthma and COPD, it is interesting to note that A_2B receptor signalling promotes the production of Th2‐type cytokines and the recruitment and degranulation of eosinophils (Belikoff et al., 2012; Karmouty‐Quintana et al., 2013a). Furthermore, it has been known for a long time that the secretion of IL‐4 by human mast cells is mediated through A_2B receptors and this increases IgE production by B cells and as a consequence triggers allergic inflammation (Ryzhov et al., 2006, 2008). In addition, there is unequivocal evidence from studies in rodent mast cells that the A₃ receptor is involved in the bronchoconstrictor response; however, this has not been verified in humans (Rudich et al., 2012). On the one hand, in primary human lung mast cells, stimulation of A₃ receptors increases IgE‐induced degranulation (Feoktistov et al., 2003; Gomez et al., 2011); and on the other, no effect was obtained in human umbilical cord blood‐derived mast cells (Hua et al., 2011) or in the LAD2 human mast cell line (Leung et al., 2014), suggesting that this response depends on the tissue studied (Gomez et al., 2011). Recently, it has been demonstrated that the primary function of the A₃ receptor is to prime the human mast cells towards tissue remodelling activity (Rudich et al., 2015). Therefore, A_2B and A₃ receptor antagonists may also be useful for the treatment of asthma and COPD (Figure 4).

Figure 3.

Modulation of wound healing, angiogenesis and fibrosis by A_2B and A₃ receptors in inflammatory cells.

Figure 4.

Deleterious effects of adenosine mediated through A_2B and A₃ receptors expressed in the lung and kidney. AAM, alternatively activated macrophages; HS, high salt; UNX, uninephrectomy.

In addition, more recently, it has been demonstrated that A_2B receptor blockade in hypoxia significantly decreased renal fibroblast proliferation and activation with a consequent reduction of profibrotic cytokine release, thus preventing the generation and development of renal fibrosis, the outcome of all types of chronic kidney disease (Tang et al., 2015). Indeed, in a novel mouse model of renal and cardiovascular disease induced through uninephrectomy and chronic high salt intake, the inhibition of A₃ receptor signalling prevented the development of hypertension and attenuated the cardiac hypertrophy and fibrosis together with renal injuries and proteinuria (Yang et al., 2016) (Figure 4). This finding accords with that of a previously reported study showing that A₃ receptor antagonism reduces the progression of renal fibrosis (Lee et al., 2013). Overall, these findings indicate that A_2B and A₃ receptor antagonists could be an interesting new therapeutic strategy for the development of new drugs to prevent tissue remodelling after tissue ischaemia and, consequently, for the treatment of fibrotic diseases including heart failure. Interestingly, it has been shown in a rat model that the blockade of A_2B receptors, beginning 1 week after myocardial infarction, reduces the pathology of this condition (Zhang et al., 2014).

Many cardiovascular diseases have atherosclerosis as an aetiological factor. In particular, endothelial dysfunction with the accumulation and oxidation of LDL leads to an increase in foam cells (Eisenstein et al., 2015). The role of adenosine receptors in the development of atherosclerosis has been investigated. Adenosine has been shown to increase the accumulation of HIF‐1α through the stimulation of all of its receptors. This is important because hypoxia is characterized by atherosclerotic plaques and HIF‐1 promotes intraplaque angiogenesis and the development of FCs (Gessi et al., 2010). The signalling mediating this effect on HIF‐1α induced by stimulation of A₁, A_2A and A_2B receptors involves ERK1/2, p38 MAPK and Akt phosphorylation, while only ERK1/2 is involved in the activation induced by A₃ receptors. In addition, A_2B and A₃ receptors stimulate VEGF production in an HIF‐1α‐dependent manner. Finally, adenosine‐stimulated FC formation was decreased by the pharmacological blockade of A₃ and A_2B receptors and by the silencing of HIF‐1α. This study indicated for the first time that A₃ and A_2B receptor antagonists may be useful for preventing the development of atherosclerotic plaques (Gessi et al., 2010). In addition, A_2B receptor antagonists have been found to inhibit fatty liver formation in mice post alcohol assumption (Peng et al., 2009). However, in contrast, in a different study, using an in vivo mouse model of atherosclerosis, the role of A_2B receptors in plaque formation was found to be protective (Koupenova et al., 2012).

Pathological role of adenosine in diabetes

It is well recognized that adenosine regulates insulin secretion, glucose homeostasis and lipid metabolism, by stimulation its receptors. The stimulation of A₁ and A_2A receptor subtypes seems to promote an antidiabetic phenotype even though recently it has been shown that blockade of A₁ receptor activation offers protection from age‐related oxidative stress and secretion of pro‐inflammatory cytokines, thus improving insulin release and effect (Yang et al., 2015). As for the A_2B receptor, its role is still controversial with a number of studies supporting a role for its agonists as a therapy for diabetes (Johnston‐Cox et al., 2012; 2014; Peleli et al., 2015). However, this protective effect is challenged by a series of studies reporting the beneficial effects of A_2B receptor antagonists. It was firstly reported that A_2B receptor antagonists behave as hypoglycaemic agents in rat models of hepatic glucose production induced by adenosine (Harada et al., 2001a,b). Furthermore, it was found that A_2B receptor activation increases glucose production by affecting glycogenolysis and gluconeogenesis in the rat liver (Yasuda et al., 2003). Following this line, A_2B receptor antagonists were shown to counteract the reduction in insulin levels induced by a non‐selective adenosine receptor agonist in pancreatic cells and plasma from rats, even though this effect is not mediated through A_2B receptor activation (Rüsing et al., 2006). A_2B receptor antagonists were also demonstrated to reduce the levels of IL‐6 and other cytokines affecting glucose and fat metabolism in a diabetic mouse model, thus improving insulin resistance (Figler et al., 2011). Furthermore, A_2B receptor blockade reduced the activation of the pro‐inflammatory caspase‐1 in rat retinal cells incubated in hyperglycaemic conditions (Trueblood et al., 2011; Vindeirinho et al., 2016). Interestingly, it has been found that high glucose levels and experimental diabetes increase the concentrations of adenosine in plasma by reducing the equilibrative NT in proximal tubule cells. This increase correlated with a marker of renal fibrosis in diabetic rats. Furthermore, the expression of profibrotic cell activation markers α‐smooth muscle actin and fibronectin was increased by stimulation of A₃ receptors (Kretschmar et al., 2016) (Figure 5).

Figure 5.

Deleterious effects of adenosine mediated through A_2B and A₃ receptors in diabetes. α‐SMA, α‐smooth muscle actin.

Therefore, before establishing a role for A_2B receptor agonists or A_2B/A₃ receptor antagonists in diabetic therapy, when looking at the evidence it is important to accurately examine the experimental conditions associated with glucose and insulin regulation including the method used for inhibiting the receptors (pharmacological vs. genetic) and the cell types or model system used (Antonioli et al., 2015; Merighi et al., 2015).

Pathological role of adenosine in cancer

Adenosine plays a role in promoting cancer development by evoking immunosuppressive effects and directly affecting the growth, metastasis and angiogenesis of tumour cells (Antonioli et al., 2014; Allard et al., 2016; Di Virgilio and Adinolfi, 2016; Ohta, 2016) (Figure 6). There is a strong relationship between cancer, hypoxia and adenosine metabolism resulting in increased levels of the autacoids in hypoxic tumours. This effect is essentially a consequence of specific alterations in the enzymes involved in adenosine production, for example, the overexpression of CD73 and the down‐regulation of adenosine kinase (ADK) are both increased by hypoxia (Ohta, 2016). Indeed, clinical studies have reported that the expression of CD73 is associated with a poor prognosis in different types of cancer (Antonioli et al., 2016), including breast (Loi et al., 2013), ovarian (Turcotte et al., 2015), prostate (Leclerc et al., 2016), brain (Quezada et al., 2013) and leukaemia (Coustan‐Smith et al., 2011; Serra et al., 2011). It has been reported that antibody‐ or genetic silencing‐mediated inactivation of CD73 inhibited breast, prostate, fibrosarcoma and melanoma tumour growth (Zhi et al., 2007; Stagg et al., 2010; 2011; 2012; Terp et al., 2013). The effects of CD73 on cancer development have been attributed to the immunosuppressive effects of A_2A and A_2B receptor stimulation (Beavis et al., 2013; Young et al., 2016). In particular, adenosine induces an immune‐tolerant micro‐environment around tumours affecting the functions of immune and inflammatory cells like T‐ and natural killer (NK) cells, macrophages and dendritic and myeloid‐derived suppressor cells (MDSC). Treg cells express high levels of CD39 resulting in an increased production of adenosine, which then inhibits the antitumour immune response; this effect of adenosine is mediated through inhibition of NK effector lymphocytes that lose their ability to recognize neoplastic cells. This effect was initially attributed to A₃ receptor activation, but is now thought to be mediated by A_2A receptors (MacKenzie et al., 1994; Ohta et al., 2012; Antonioli et al., 2013). As for tumour‐associated macrophages, stimulation of A_2A and A_2B receptors induces an enhanced expression of the M2 phenotype, which contributes to the angiogenesis, proliferation and metastasis of tumour cells (Csóka et al., 2012). In addition, A_2B receptors mediate the increased production of VEGF by dendritic cells; VEGF is essential for tumour angiogenesis and growth, and favours the activity of MDSC that are able to suppress immune surveillance (Ryzhov et al., 2011). Overall, these results have led to the development of phase I clinical trials that are now underway to consider the safety and the efficacy of CD73 and A_2A receptor inhibitors in cancer patients (Allard et al., 2016).

Figure 6.

Deleterious effects of adenosine mediated through A_2A and A_2B receptors in cancer.

The adenosine machinery also mediates its effects directly in tumour cells by affecting their proliferation and cell death through recruitment of different receptors. In particular, opposite effects on cell growth and motility have been observed, with A₁, A_2A and A_2B receptors being promoters whilst A₃ receptors are inhibitors of cell proliferation (Merighi et al., 2002; 2005; Gessi et al., 2011a; Antonioli et al., 2013; Borea et al., 2015; Ledderose et al., 2016). Indeed, it is important to note that clinical trials are undergoing for A₃ receptor agonists, based on their antiproliferative effects, representing an opportunity for new anticancer drugs (Borea et al., 2016).

Nevertheless, it has been reported that A_2B receptors have a prometastatic and prosurvival effect and that stimulation of A₁, A_2A and A₃ receptors increases melanoma growth, neovascularization, angiogenesis and macrophage infiltration in CD73 KO mice (Linden, 2013; Ntantie et al., 2013; Koszałka et al., 2016) (Figure 5).

Interestingly, an overexpression of A₃ receptors has been reported in different tumours including colon, breast, hepatocellular and mesothelioma (Madi et al., 2004; Gessi et al., 2004; 2011b; Varani et al., 2011; Borea et al., 2015). Furthermore, an up‐regulation of A_2B receptors has also been observed in colorectal cancer (Ma et al., 2010). Therefore, we can conclude that the important effects adenosine exerts on the progression and development cancer depend on the subtype of adenosine receptor expressed in each tumour.

The negative actions of long‐lasting increases in the extracellular adenosine concentration lead to overactivation of membrane receptors, thus opening the way for the clinical development of specific antagonists. First of all, starting with the CNS injuries, A_2A selective antagonists may have a role in the therapy of ischaemia, where a careful attention to administration time and to the dose of A_2A receptor antagonist has to be considered for a novel therapeutic strategy. PD and other neurodegenerative diseases such as HD and AD, major depression, schizophrenia, epilepsy, acute and chronic stress, RLS and memory fear are pathologies for which A_2A selective inhibitors could became new drugs. Even though istradefylline, in combination with L‐DOPA, is now approved in Japan for reducing the “off time” in PD, to date no disease‐modifying treatment is available. Therefore, the research to identify new therapeutics for PD should be increased and the reasons why the translation of basic research to disease‐modifying therapies has been unsuccessful so far should be investigated. The new elegant evidence regarding the modulation of heterotetrameric structure of the A_2A/D₂ receptor complex and the similarity between A_2A receptor agonist and antagonist effect on the affinity and intrinsic efficacy of D₂ receptor ligands offers a novel model that can provide new clues about how to adjust the clinical use of these drugs (Ferré et al., 2016).

In MS, an autoimmune disease of central origin, both A_2A and A_2B receptor antagonists attenuated the increased inflammation and then the development of its pathology. Hence, the blockade of these receptors might be a new strategy for the development of new drugs for MS therapy.

Understanding the effect of A_2B receptor modulation in intestinal inflammation, glucose and lipid regulation is necessary to clarify whether this subtype is protective or injurious in both colitis, diabetes and atherosclerosis, respectively, with the aim of adequately recognizing its therapeutic potential.

Furthermore, A_2B and A₃ receptor antagonists may represent a starting point for the development of novel drugs to hamper remodelling occurring after tissue ischaemia and for the therapy of consequent fibrotic diseases arising in the lung, kidney and heart.

The pioneering research by Sitkowsky and co‐workers on tumours and hypoxia also opens the way for the use of A_2A receptor antagonists in the therapy of cancer. Due to their immunosuppressive effects and tumour tolerant behaviour, now both A_2A an A_2B receptors are considered important and promising targets for the development of a novel class of antitumoural drugs able to weaken the hypoxia‐adenosine mediated signalling pathways. It is important to note that existing A_2A receptor antagonists tested in human clinical studies for PD have already demonstrated their safety profile. However, in the future, it will be important to consider the development of new A_2A receptor blockers unable to cross the BBB, thus avoiding the occurrence of potential neurological side effects in cancer patients with non‐cerebral tumours (Hatfield and Sitkovsky, 2016). This task may be supported by important advances in the structure‐based design of A_2A receptor antagonists that take advantage of the recently revealed molecular basis of this adenosine subtype (Carpenter et al., 2016; Ye et al., 2016; Jazayeri et al., 2017).

Conclusions and perspectives

It is well known that adenosine is beneficial, adopting the role of guardian angel when present at physiological levels or when increased in acute situations (Borea et al., 2016). However, we have to remember that there are instances in which its chronic overproduction is pathological. In this review, we have presented evidence for its bad side, particularly in association with ischaemia, neurodegenerative diseases, fibrosis, hepatic steatosis, colitis, asthma, diabetes and cancer, while focusing on the potentiality of this system as a new therapeutic weapon in the battle against important unmet diseases. Indeed, A_2A, A_2B and A₃ receptor antagonists may find a specific/unique place in different clinical applications. Overall, this is an exciting period for scientists involved in the adenosinergic field seeking to harness the information for new therapeutic applications.

Conflict of interest

The authors declare no conflicts of interest.

Borea, P. A. , Gessi, S. , Merighi, S. , Vincenzi, F. , and Varani, K. (2017) Pathological overproduction: the bad side of adenosine. British Journal of Pharmacology, 174: 1945–1960. doi: 10.1111/bph.13763.