Adenosine A_2A receptor antagonists: from caffeine to selective non-xanthines

Abstract

A long evolution of knowledge of the psychostimulant caffeine led in the 1960s to another purine natural product, adenosine and its A_2A receptor. Adenosine is a short-lived autocrine/paracrine mediator that acts pharmacologically at four different adenosine receptors in a manner opposite to the pan-antagonist caffeine and serves as an endogenous allostatic regulator. Although detrimental in the developing brain, caffeine appears to be cerebroprotective in aging. Moderate caffeine consumption in adults, except in pregnancy, may also provide benefit in pain, diabetes, and kidney and liver disorders. Inhibition of A_2A receptors is one of caffeine’s principal effects and we now understand this interaction at the atomic level. The A_2A receptor has become a prototypical example of utilizing high-resolution structures of GPCRs for the rational design of chemically diverse drug molecules. The previous focus on discovery of selective A_2A receptor antagonists for neurodegenerative diseases has expanded to include immunotherapy for cancer, and clinical trials have ensued.

1 |. INTRODUCTION

Caffeine (1,3,7-trimethylxanthine; Table 1, compound 1), found in coffee, tea, cola, chocolate, and other foods, is the most widely ingested drug substance (Reyes & Cornelis, 2018). The boost in focus, alertness, and enhanced mood from a cup of morning coffee is essential for many in the Western world. Annual consumption of caffeine-containing beverages average 348 L per capita in North America, 200 L per capita in Europe, although less in other regions (Reyes & Cornelis, 2018). Coffee contains dozens of pharmacologically active substances, among them are the alkylxanthines, principally caffeine, and antioxidant polyphenols such as the catechol derivative caffeic acid and its esters (dePaula & Farah, 2019; Fischer, Victor, Robinson, Farah, & Martin, 2019). The actions and toxicology of caffeine, first isolated in 1820, have been extensively investigated (Fischer et al., 2019; Fredholm, Bättig, Holmén, Nehlig, & Zvartau, 1999; Temple et al., 2017). Caffeine is not an addictive drug, but habituation and withdrawal can occur. Caffeine withdrawal can cause rebound throbbing headaches, drowsiness, depressed mood, fatigue, and anxiety (Fredholm et al., 1999). Excessive intake has negative consequences, such as tremors and tachycardia, but daily doses of <400 mg have no serious adverse effects in healthy adults (Reyes & Cornelis, 2018). In fact, there may be incidental medical benefit of caffeine intake in some conditions (Bravi et al., 2007; Chen, 2019; Dranoff, 2018; Hu et al., 2018; Neves et al., 2018; Ramamoorthy et al., 2017). For centuries, caffeine has been casually consumed in various forms, but in order to truly understand its effects, we first need to explore its pharmacology and mechanism of action. Is coffee a medicinally valuable tool worthy of study?

TABLE 1.

Affinity of caffeine and other simple alkylxanthines at adenosine receptors (A₁AR, A_2AAR, A_2BAR and A₃AR) and at other protein targets

Derivative	A1AR	A2AAR	A2BAR	A3AR	Other targetsa
1, Caffeine	10.7	9.56	10.4	13.3	PDE1b 480, PDE2 710, PDE3 > 100, PDE4 > 100, PDE5 690, GABAA 500, MAO-A 761, ryanodine receptor
2, Theophylline	6.77	6.7	9.07	22.3	PDE1b 280, PDE2 270, PDE3 380, PDE4 150, PDE5 630, PI3-kinase 100, HDAC
3, Paraxanthine	21 (r)	19.4 (r)	4.5	>100 (r)	cGMP-preferring PDE
4, Theobromine	105 (r)	>250 (r)	130	>100 (r)	PDE4, poly (ADPribose) polymerase-1,160
5, Enprofylline	160	32	4.73	65	PDE1b 270, PDE2 260, PDE3 110, PDE4 100, PDE5 590

Data shown are K_i or IC₅₀ (μM) for human receptors, unless otherwise noted. Caffeine, theophylline, and paraxanthines, but not theobromine, are natural products and potent behavioural stimulants. Enprofylline is a synthetic, not naturally occurring, alkylxanthine. The numbers in bold refer to the corresponding chemical structures shown in Figure 1.

^aReferences: Daly, 2007; Foukas et al., 2002; Hirano et al., 2006; Ito et al., 2002; Jacobson, IJzerman, & Linden, 1999; Ukena, Schudt, & Sybrecht, 1993; Monteiro, Alves, Oliveira, & Silva, 2016; Orrú et al., 2013; Petzer, Pienaar, & Petzer, 2013.

There is a long history of exploring the pharmacological and physiological effects of caffeine in the CNS and systemically. Early studies led to the identification of the action of caffeine as an antagonist of adenosine and specifically through the adenosine A_2A receptor as one of the principal targets (Fredholm et al., 1999). The A_2A receptor is one of four receptor subtypes that respond to extracellular adenosine as an endogenous regulator of many physiological processes. Furthermore, numerous X-ray and electron cryo-microscopy (cryo-EM) structures of the A_2A receptor have been reported, which have enabled the rational design of novel ligands of this receptor, both antagonists and agonists, and an atomistic understanding of its activation (Doré et al., 2011; Eddy, Lee, et al., 2018; García-Nafría, Lee, Bai, Carpenter, & Tate, 2018; Lebon et al., 2011; Xu et al., 2011). New antagonists targeting this receptor are being developed for Parkinson’s disease (PD) and other neurodegenerative conditions, as well as for the treatment of cancer, while agonists have been proposed for reducing inflammation (Jacobson, Tosh, Jain, & Gao, 2019; Merighi et al., 2019; Sorrentino et al., 2019). This review traces research on A_2A receptors antagonists from its origins in alkylxanthine natural products to highly potent and selective non-xanthine antagonists.

2 |. NATURALLY OCCURRING LKYLXANTHINES AS THE PROTOTYPICAL ADENOSINE RECEPTOR ANTAGONISTS AND THEIR THERAPEUTIC USE

The pharmacology, absorption, and metabolism of plant-derived alkylxanthines in humans have been well characterized (Fredholm et al., 1999; Temple et al., 2017). Simple alkylxanthines, including caffeine and its main metabolite paraxanthine (Table 1, compound 2), and its close analogues theophylline (Table 1, compound 3) and theobromine (only trace amounts in coffee; Table 1, compound 4), can act non-selectively at the four adenosine receptor subtypes (Table 1). Paraxanthine is roughly equipotent to caffeine, but theobromine, found in chocolate, is weaker at the adenosine receptors. Caffeine not only acts at A_2A receptors but also has affinity for other adenosine receptor subtypes in humans. For example, A₁ receptor antagonism by caffeine causes noticeable effects such as diuresis and tachycardia. Antagonism of this receptor is likely to be the basis for some of the CNS effects of caffeine, such as the effects on hippocampal synaptic transmission (Chen, 2019; Fang et al., 2017; Lopes, Pliássova, & Cunha, 2019); while both A₁ and A_2A receptors are involved in sleep (Lazarus, Chen, Huang, Urade, & Fredholm, 2019). Thus, caffeine can be considered a pan-antagonist at the human adenosine receptors, although it is much weaker at rodent A₃ receptors. The half-life of caffeine is variable but typically 4–5 h in the healthy adult. Due to its longer half-life in young children, limiting intake is important. Caffeine readily distributes throughout the body, including the brain, and its exposure in humans is generally independent of the type of drink (concentration-normalized), its temperature, or rapidity of administration (White et al., 2016). Peak plasma caffeine concentrations of 3–4 μg·ml⁻¹, that is, ~20 μM, are achieved upon drinking a single coffee beverage (160 mg).

Human genetic variation can alter the effects of caffeine, leading to a wide range in caffeine tolerance and stimulant responses. The liver enzyme mostly responsible for caffeine metabolism, by oxidative removal of its 3-methyl group, is CYP1A2, which is genetically heterogeneous within populations (dePaula & Farah, 2019). Substitution of a thymine (T) for cytosine (C) in SNP rs2472297, which affects ~10% of humans, is associated with higher coffee consumption. Other CYP1A2 SNPs can affect the rate of caffeine metabolism. Similarly, those ingesting >400 mg·day⁻¹ caffeine tend to have certain variants of the aryl hydrocarbon receptor (AHR) gene, for example, C allele of the SNP rs4410790 (Josse, Da Costa, Campos, & El-Sohemy, 2012). The rate of N-3 demethylation of caffeine is dependent on this rs4410790 allele, and a ¹³C-breath test administered after ingesting isotopically labelled caffeine is indicative of its gene polymorphism (Ishii, Ishii, Nakayama, Takahashi, & Asai, 2020). Variations in the gene ADORA2A coding for A_2A receptor protein are associated with sensitivity and responses to caffeine. For example, people with a C allele in the most widely studied ADORA2A SNP, rs5751876, are more sensitive to caffeine’s sleep effects (Huin et al., 2019). Its TT genotype is associated with increased anxiety after consuming caffeine.

Alkylxanthines have long been used in human therapeutics, and new applications are continually being explored. There are 854 studies listed in ClinicalTrials.gov using the search term “caffeine” (accessed November 25, 2019). Theophylline is an inhibitor of phosphodiesterases (PDEs) and a slightly more potent adenosine receptor antagonist than caffeine. Beginning in the 1920s, the use of theophylline became common for treating asthma and chronic obstructive pulmonary disease (COPD) (Barnes, 2010; Schultze-Werninghaus & Meier-Sydow, 1982). The anti-asthmatic effect of the alkylxanthines might be due to blockade of the A_2B receptor subtype, but this is controversial as there are also beneficial effects of the activation of this receptor, in asthma (Gao & Jacobson, 2017). Other alkylxanthines used clinically for pulmonary conditions have included enprofylline (3-propylxanthine; Table 1, compound 5), which weakly but selectively antagonizes the A_2B receptors.

There are numerous reports, often contradictory, of other health effects of caffeine, for example, its reported benefits in liver disease, diabetes, cancer, and CNS disorders. Caffeine is proposed to protect against fibrosis and other effects of ethanol in the liver (dePaula & Farah, 2019; Dranoff, 2018). Caffeine has been used to improve exercise performance and weight loss, and excessive levels in athletes can be monitored. Included in over-the-counter medications, caffeine relieves headaches, which is likely to be due to blocking the vascular A_2A receptors to prevent meningeal arterial dilation (Haanes et al., 2018). However, mixed A₁ / A_2A receptor-selective antagonists were more effective than selective A_2A receptor antagonists, such as JNJ-41501798 (Figure 1, compound 17), in a rat migraine model. In AIDS patients, caffeine consumption appeared to increase CD4 T cell counts and decrease HIV viral load, possibly due to its potential to boost immunity via A_2A receptor antagonism (Ramamoorthy et al., 2017). In the gastrointestinal system, caffeine elevates the peptides gastrin and cholecystokinin, and consequently coffee stimulates colonic motility in some individuals. Chronic caffeine administration enhanced Cl⁻ secretion into the rat intestine, resulting in an anti-hypertensive effect in salt-sensitive rats (Wei et al., 2018).

FIGURE 1.

This Figure shows the structures (compounds 6 – 21) of some synthetic potent and selective heterocyclic A_2A receptor antagonists. When radiolabelled, an asterisk (*) indicates the site of isotopic labelling. The structures of simple alkylxanthines (compounds 1 – 5) are presented in Table 1

Caffeine also plays a neuroprotective role. Caffeine reduces memory deterioration in the aging brain, with A_2A receptor antagonism being a mechanism responsible for caffeine’s beneficial effect on hippocampal long term-potentiation (LTP) associated with memory (Lopes et al., 2019; Temido-Ferreira et al., 2018). The negative correlation between PD and caffeine intake is reasonably well established, although interventional studies are not definitive (Ross et al., 2000). There is also an epidemiological inverse correlation of caffeine ingestion with Alzheimer’s disease (AD) (Cellai et al., 2018; Cunha, 2016). Although there is retrospective evidence of an association of caffeine with lower incidence of certain neurodegenerative diseases, a prospective clinical study with evidence of caffeine’s neuroprotection is lacking. However, caffeine consumption is associated with an early onset of Huntington’s disease (Simonin et al., 2013), and its effects in multiple sclerosis or amyotrophic lateral sclerosis are unclear. One study found that caffeine can detrimentally affect behavioural impairments in AD (Baeta-Corral, Johansson, & Giménez-Llort, 2018). Caffeine consumption has a complex relationship with ethanol abuse and addiction (Sanmiguel, López-Cruz, Müller, Salamone, & Correa, 2019). In mice, caffeine increased ethanol intake in moderate consumers having unlimited access, but caffeine reduced it when ethanol was removed and reintroduced. A high dose of a more selective A_2A receptor antagonist prodrug MSX-3 (Figure 1, compound 10) also increased ethanol administration. Caffeine was previously tested in children as an adenosine receptor antagonist to treat attention-deficit hyperactivity disorder (ADHD) and abandoned because of side effects, but its use might be reconsidered (França, Takahashi, Cunha, & Prediger, 2018; Ioannidis, Chamberlain, & Müller, 2014). By antagonizing the A_2A receptors, caffeine blocks the ethanol-induced increase in non-REM sleep (Fang et al., 2017).

Some but not all the effects of coffee drinking are ascribed to adenosine receptor antagonism. Perhaps the most puzzling human health correlation with increased coffee consumption is with reduced overall mortality in European countries (Freedman, Park, Abnet, Hollenbeck, & Sinha, 2012; Gunter et al., 2017). To try to explain the correlational advantage of coffee intake, a clinical survey of plasma biomarkers associated with chronic inflammatory diseases reported favourable reductions in coffee drinkers, but it was unexplained why even decaffeinated coffee drinkers showed a benefit (Hang et al., 2019). In a separate clinical analysis, caffeine consumption did not correlate with lower cardiovascular or cancer mortality but had a lower risk in females for diabetes and for all-cause mortality (Neves et al., 2018). A prospective study of human consumption of non-decaffeinated coffee indicated a lower risk of chronic kidney disease (Hu et al., 2018). In a meta-analysis, caffeine consumption was associated with a lower risk of hepatocellular carcinoma and other cancers (Bravi et al., 2007; dePaula & Farah, 2019).

Adenosine acting at adenosine receptors provides an adaptive advantage for organisms—so we need to determine the risks of blocking these receptors with caffeine, theophylline, and other adenosine receptor antagonists: Common undesired effects ascribed to caffeine include anxiety, insomnia, fatigue, gastroesophageal reflux, and hypertension. With huge doses of caffeine, tremors, vomiting, gastrointestinal irritation, convulsions, and arrhythmias (presumably by blocking the A₁ receptor), and death (likely resulting from arrhythmias) can occur (Cappelletti et al., 2018; dePaula & Farah, 2019; Voskoboinik, Kalman, & Kistler, 2018). Diuresis could be considered a minor negative consequence of caffeine intake. During pregnancy, caffeine use is strongly discouraged as it may interfere with the formation of neuronal connectivity in the fetal brain and is similarly risky in the young developing brain (Atik et al., 2017; Silva et al., 2013). Chronic caffeine treatment has been noted to either up- or down-regulate the expression levels of many other non-adenosine receptors in the mouse brain, so there might be implications for various signalling systems in the human brain as well (Shi, Nikodijevic, Jacobson, & Daly, 1993). Adenosine receptor agonists display benefit in animal models of inflammatory and ischaemic conditions, and agonists of A_2A receptors, such as regadenoson (used clinically as a myocardial perfusion imaging agent), and other adenosine receptors have been in clinical trials for sickle cell disease and other maladies. Do caffeine and other A_2A receptors antagonists therefore have a liability due to their charging of the innate immune system, for example, to worsen chronic inflammation? There are contradictory reports indicating either salutary or harmful effects of human caffeine consumption in inflammatory states, or in combination with drugs, such as the adenosine-boosting drug methotrexate for rheumatoid arthritis (Furman et al., 2017; Malaviya, 2017).

The general importance of adenosine in health and pathology has been explored using mouse lines in which one or more of the adenosine receptors has been genetically deleted. The potentially good news for healthy caffeine drinkers is that the lack of each of the receptors generally does not lead to strong disease conditions in these knockout (KO) mice, with some exceptions such as arthritic symptoms in aging A_2A receptor-KO mice. In fact, the role of endogenous adenosine was recently explored using quadruple adenosine receptor-KO mice (Xiao, Liu, Jacobson, Gavrilova, & Reitman, 2019). These rodents completely lack typical adenosine pharmacology, such as an adenosine receptor agonist-induced reduction of core body temperature. Furthermore, they lack a behavioural stimulant response to caffeine, again emphasizing that the adenosine receptors are an important mechanism of action of caffeine. The absence of a pronounced phenotype in this mouse line indicated that adenosine is to be considered an endogenous allostatic, rather than homeostatic, regulator in the body. Thus, the adenosine receptors serve to correct an imbalance induced by physiological stress, such as hypoxia, and caffeine might interfere with this endogenous protective mechanism in such disease states.

3 |. IDENTIFICATION AND CHARACTERIZATION OF THE A_2A RECEPTOR AS A MOLECULAR TARGET OF CAFFEINE

Caffeine is also associated with effects on other intracellular biochemical pathways (Table 1), principally: (a) inhibition of PDEs (typically >100 μM), to increase cAMP; (b) a rise in cytosolic calcium concentration (typically >500 μM), due to blocking Ca²⁺ reuptake by the sarcoplasmic reticulum; (c) activation of the ryanodine receptor with roughly mM affinity to release Ca²⁺ stores (Porta et al., 2011; Ukena et al., 1993). Historically, the action of caffeine at adenosine receptors has been neglected as a mechanism in favour of PDE inhibition, as is written in some medical school textbooks (Daly, 2007). In addition, caffeine is also a weak (mM) inhibitor of monoamine oxidase A (MAO-A) and has served as a scaffold for more potent MAO-A inhibitors (Petzer et al., 2013). Theophylline (~10 μM) induces histone deacetylase (HDAC) in bronchoalveolar macrophages by a non-adenosine receptor mechanism, to inhibit inflammatory gene transcription (Ito et al., 2002). However, caffeine decreased HDAC immunoreactivity in the hippocampus and striatum of dopamine-depleted rats, which might contribute to its neuroprotection (Machado-Filho et al., 2014). Caffeine impeded the progression of quiescent mouse epidermal cells into the cell cycle with ED₅₀ ~ 0.7 mM, an effect that might limit in vivo carcinogenesis (Hashimoto et al., 2004). Nevertheless, at doses normally ingested, the most relevant mechanism of action of caffeine is adenosine receptor antagonism (Daly, 2007; Fredholm et al., 1999).

Historically, the effects of both caffeine and adenosine on the cardiovascular system, including heart rate and blood vessel dilation, were carefully studied in the 1920s (Drury & Szent-Gyorgyi, 1929; Heathcote, 1920). However, caffeine’s effect on adenosine-mediated depressant functions was not extensively explored until the 1950s and later (De Gubareff & Sleator, 1965; Guthrie & Nayler, 1967; Ther, Muschaweck, & Hergott, 1957). Nichols and Walaszek (1963) reported caffeine’s antagonism of the vasodepressor effect of adenosine in the chicken, rabbit, cat, and dog. De Gubareff and Sleator (1965) found that caffeine induced calcium release, and it was more potent in antagonizing adenosine-mediated than cholinergic-mediated contractions of both human and guinea pig isolated atria.

In the meantime, Sutherland and Rall (1958) identified caffeine as a PDE inhibitor. By inhibiting PDE, caffeine was able to boost cAMP production induced by noradrenaline. Shimizu, Daly, and Creveling (1969) reported that both histamine and adenosine induce cAMP accumulation in guinea pig cerebral cortical slices, and the PDE inhibitor caffeine greatly increased histamine-induced but slightly decreased adenosine-induced cAMP production, suggesting a possible inhibitory effect on both PDE and adenosine. The finding from Shimizu et al. (1969) is in line with a report by Rall and Sattin (1968) that another more potent PDE inhibitor, theophylline, was able to block instead of enhancing adenosine-induced cAMP accumulation. It is noted that the concentrations of both adenosine and caffeine used in earlier studies are very high (1 mM or higher). Sattin and Rall (1970) found that 0.1-mM caffeine already produced a substantial effect, and 0.5-mM caffeine was able to completely block cAMP accumulation induced by adenosine (50 μM), which clearly demonstrated an inhibitory rather than an enhancing effect of caffeine on cAMP accumulation. Subsequently, caffeine, theophylline, and other methylxanthines were established as adenosine receptor antagonists (Bruns, Lu, & Pugsley, 1986; Fredholm & Persson, 1982; Londos, Cooper, & Wolff, 1980; Snyder, Katims, Annau, Bruns, & Daly, 1981; van Calker, Müller, & Hamprecht, 1979). Although it has been suggested earlier that the behavioral stimulatory effect of caffeine is mainly via the A_2A receptors, Ledent et al. (1997) demonstrated that A_2A receptor knockout mice were viable and bred normally, but caffeine, which normally stimulates exploratory activities, became a depressant of exploratory behaviour. Radioligand binding demonstrated the micromolar A_2A receptor affinity of caffeine (Table 1). The A_2A receptor was first identified as a protein distinct from the A₁ receptor using photoaffinity labelling of the receptor in striatal membranes using a radioiodinated azide-derivatized selective A_2A receptor agonist (Barrington, Jacobson, Hutchison, Williams, & Stiles, 1989). The A_2A receptor was initially cloned as an orphan receptor and deorphanized in ~1990 (Libert et al., 1989; Maenhaut et al., 1990).

4 |. PRECLINICAL AND CLINICAL RESULTS WITH POTENT A_2A RECEPTOR NTAGONISTS

Many potent and selective A_2A receptor antagonists, initially xanthines (Figure 1, compounds 6–9) and more recently non-xanthine heterocycles (Figure 1, compounds 10–19) have been reported, with varying degrees of selectivity (Table 2). One of the chemical challenges in developing A_2A receptor antagonists has been the substitution of the furan group, present in ZM241385 (Figure 1, compound 10) and many other non-xanthine antagonists (Figure 1, compounds 11, 12, 14, and 19), that might have a higher chance of yielding reactive metabolites. A_2A receptor antagonists are of interest for the treatment of neurodegenerative diseases and as combined therapy or monotherapy in the immunotherapy of cancer. In addition, there might be potential for the use of A_2A receptor antagonists in the liver to prevent cirrhosis (Dranoff, 2018).

TABLE 2.

Binding affinity of diverse antagonists at human adenosine receptors

Compound	A1ARa	A2AARa	A2BARa	A3ARa
	pKi
6 XAC	8.17,8.92b	7.74,7.20b	8.11	7.59
7 istradefyllinec	5.55	7.44	5.74	<5.52
8 CSC	4.55b	7.27b	5.09	<5b
9a MSX-2	5.60	8.10	<5	<5
10 ZM241,385	6.11	8.80	7.12	6.13
11 Preladenant	5.83	8.96	<5.77	<6
13 SYN-115	5.87	8.30	6.15	5.80
14 Vipadenant	7.19	8.89	7.20	6.00
15 ANR94	5.62	7.32	<4.52	4.68
16 CPI-444	6.75	8.45	5.82	5.61
17 JNJ-41501798	5.10	7.94	n.d.	n.d.
18 HTL-1071	6.80	8.77	7.19	<5
19 Compound 504	6.99	9.93	7.47	6.37
20 PBF-509	5.60	7.92	6.00	5.30
21 AB928	n.d.	8.85	8.70	n.d.

The numbers in bold refer to the corresponding chemical structures shown in Figure 1. Abbreviation: n.d., not determined or not reported.

^aMean pK_i or pIC₅₀ μM, radioligand binding.

^bRat adenosine receptors.

^cKW-6002.

The A₁ and A_2A receptors are highly expressed in the brain, although A_2A receptors hves more limited distribution, being highly expressly in the striatum (on GABAergic enkephalinergic striatopallidal neurons) and the olfactory tubercle (Cunha, 2016; Rodrigues, Marques, & Cunha, 2019). The A_2A receptors are up-regulated in models of brain dysfunction, damage, and aging. Adenosine negatively regulates dopaminergic signalling in the striatum. This is thought to involve a direct interaction between the receptors, that is, a A_2A receptor-dopamine D₂ receptor heterodimer, as well as neural circuits (Borroto-Escuela & Fuxe, 2019). Therefore, adenosine antagonists can allosterically increase the effects of residual dopamine through a striatal heterodimer. A_2A receptor antagonists increase dopaminergic signalling by blocking the basal ganglia indirect pathway, which explains the symptomatic relief in PD (Jazayeri, Andrews, & Marshall, 2017). A_2A receptor antagonists decrease the off-time and dyskinesias arising from prolonged use of L-DOPA. Also, antagonism of presynaptic A₁ receptors by caffeine disinhibits the release of various other transmitters, and internally negatively modulating A₁-A_2A receptor heterodimers have also been detected that are a separate pharmacological target (Borroto-Escuela & Fuxe, 2019; Gonçalves et al., 2019; Hinz et al., 2018). Inhibitory A₁ and facilitatory A_2A receptors modulate the activity of excitatory glutamatergic synapses. The neuroprotective effects of A_2A receptor antagonists may reflect the reduction of glutamate release and oxidative stress or increased integrity of the blood brain barrier and release of the inhibitory transmitter GABA (Fredholm et al., 1999; Hurtado-Alvarado, Domínguez-Salazar, Velázquez-Moctezuma, & Gómez-González, 2016). Thus, the effects of caffeine throughout the brain as an adenosine receptor antagonist are manifold.

Moreover, there might be a cerebroprotective function and cognitive benefit of A_2A receptor antagonists, in addition to the role they play in relieving in the motor deficits of PD (Chen & Cunha, 2020). Potent A_2A receptor antagonists, like caffeine, appear to be beneficial for the aging brain. Mechanistic probing of neuroprotection by A_2A receptor antagonists suggest that blocking both neuronal and non-neuronal (e.g., glial cell) A_2A receptors preserve brain function in in vivo degenerative models (Cunha, 2016). However, in the developing brain, activation of A_2A receptors is required to form proper connectivity (Silva et al., 2013) and, consequently, its antagonism can be detrimental (Rodrigues et al., 2019). A_2A receptors are up-regulated in the aging and stressed brain, in AD and in a Tauopathy model where it potentiates memory deficits (Carvalho et al., 2019; Kaster et al., 2015; Orr et al., 2015), and A_2A receptors are activated in amyloidogenesis (Rodrigues et al., 2019; Viana da Silva et al., 2016). Its blockade ameliorates memory dysfunction in a triple transgenic mouse model and other in vivo models of AD (Lopes et al., 2019; Silva et al., 2018). Blocking A_2A receptors suppresses microglial reactivity and consequently neuroinflammation (Madeira, Boia, Ambrósio, & Santiago, 2017). Caffeine and other A_2A receptor antagonists or A_2A receptor KO have distinct effects in the brain to reduce pathological Tau proteins and the C1q complement system in microglia (Carvalho et al., 2019; Cellai et al., 2018; Orr et al., 2015). Neuronal A_2A receptor overexpression exacerbates memory and learning impairment and increases gene expression connected to microglial function. Selective A_2A receptor antagonists display antidepressant activity in animal models, including increasing operant behaviour response rates (Bartoli, Burnstock, Crocamo, & Carrà, 2020). Caffeine also reduces depressive-like symptoms in PD patients, and it is suggested that selective A_2A receptor antagonists might ultimately have a role in normalizing mood (Chen & Cunha, 2020). Thus, blockade of the A_2A receptor by caffeine or selective A_2A receptor antagonists provides benefit to humans and in animal models related to PD, AD, impairment from traumatic brain injury, ADHD, risk to develop stroke, depression, and suicidal behaviour (Cunha, 2016).

Imaging adenosine receptors in the brain could potentially be used routinely in the clinic to establish the occupancy of the A_2A receptors upon administration of low MW drugs and to probe levels of endogenous adenosine. PET using [¹⁸F] or [¹¹C] A_2A receptor antagonists, such as [¹¹C]preladenant (Figure 1, compound 11), has been established as a means of imaging the receptor in the brain, both in experimental animals and human subjects (van Waarde et al., 2018). [¹¹C]SCH442416 (structure not shown), an A_2A receptor antagonist, was used to establish the human brain receptor occupancy of a potential anti-PD drug vipadenant (Figure 1, compound 14 ; Brooks et al., 2010). Single-photon emission computerized tomography (SPECT) using an labelled A_2A receptor antagonist [¹²³I]MNI-420, a pyrazolo[4,3-e][1,2,4]triazolo[1,5-c]pyrimidin-5-amine (Figure 1, compound 12), has also been reported (Vala et al., 2016). However, endogenous adenosine levels were unable to be determined using rat PET imaging using adenosine receptor antagonist tracers.

Selective A_2A receptor antagonists may have benefit in pain treatment, as suggested in the original A_2A receptor knockout mouse study (Ledent et al., 1997; Varano et al., 2016). An A_2A receptor antagonist reported by Varano et al. was shown effective in acute pain through its peripheral action (Varano et al., 2016). The vasodilatory action of A_2A receptor antagonists in the trigeminovascular system reduce migraine pain (Haanes et al., 2018).

The potential use of A_2A receptor antagonists in treating cancer was pioneered by Sitkovsky and coworkers, and several clinical trials are in progress (Fong et al., 2020; Kjaergaard, Hatfield, Jones, Ohta, & Sitkovsky, 2018). Adenosine levels and adenosinergic signalling are elevated in the tumour micro-environment (TME). This signalling shifts T-cells from the aggressive effector (T_eff) state to the immunosuppressive regulatory (T_reg) state. Macrophage polarization in the TME tends to convert the M1 activated state to the M2 anti-inflammatory state. A_2A receptor activation suppresses the maturation and proliferation of natural killer (NK) cells, and blocking the receptor in NKs promotes their anti-tumour activity (Young et al., 2018). Thus, blocking the formation of adenosine or its signalling in the TME is potentially useful clinically in cancer treatment, particularly in combination with immunotherapy. Therefore, selective A_2A receptor antagonists—or those with mixed A_2A/A_2B receptor selectivity—are actively being considered for cancer treatment, either as monotherapy or as cotherapy with other immunotherapy agents, such as anti-PD-1 and anti-CTLA4 antibodies (Fong et al., 2020).

Many A_2A receptor antagonists of diverse chemotype have been reported. Among the earliest, istradefylline (Nouriast; Figure 1, compound 7), an 8-styrylxanthine analogue of caffeine, was recently approved by the US Food and Drug Administration (FDA) (August 28, 2019) as co-therapy with levodopa/carbidopa in adult PD patients to reduce “off” episodes (Chen & Cunha, 2020). It was first approved for use in Japan (2013), and additional evidence from clinical trials and post-marketing surveillance has accrued for its clinical efficacy to reduce off-time, mood alterations, and other symptoms in PD patients receiving L-DOPA (Chen & Cunha, 2020; Takahashi, Fujita, Asai, Saki, & Mori, 2018). In dilute solutions, 8-styrylxanthines are photosensitive with an isomerization of the styryl group for trans to cis. Thus, their experimental use requires attention to lability in light of aqueous solutions for injection. Other A_2A receptor-selective styrylxanthines that have served as pharmacological probes include CSC (Figure 1, compound 8) and its phosphate prodrug for increased aqueous solubility (Jacobson & Müller, 2016).

Various drug-like non-xanthine heterocycles have been discovered as high affinity A_2A receptor antagonists for use in the CNS or in cancer immunotherapy. Both pharmacodynamic and pharmacokinetic factors are considered in selecting clinical candidate antagonists (Congreve et al., 2012). The factors involved in success or failure of A_2A receptor antagonists in PD clinical trials were recently reviewed (Chen & Cunha, 2020). Clinical trials of A_2A receptor antagonists for cancer therapy in combination with diverse immunotherapies are ongoing for a wide variety of tumour types (Merighi et al., 2019). Preladenant (Figure 1, compound 11) that lacked significant efficacy in Phase III clinical trials for PD, alone or combined with L-DOPA (Chen & Cunha, 2020), is being repurposed for possible cancer application. Recently introduced potent triazene A_2A receptor antagonists, such as HTL-1071, which is currently designated as AZD4365, a 1,2,4-triazin-3-amine, (Figure 1, compound 18), were initially proposed for treatment of ADHD and are now being developed for cancer (Jazayeri et al., 2017). Compound 504 (Figure 1, compound 19) was shown to have favourable pharmacokinetic properties and efficacy in ADHD models. An A_2A receptor antagonist may provide benefit in behavioural pathologies related to social function, including anxiety and motivation (López-Cruz et al., 2017). Recent Phase I results in treatment-refractory renal cell cancer with ciforadenant (CPI-444, formerly V-81444, a triazolo[4,5-d]pyrimidin-5-amine, (Figure 1, compound 16) indicated enhanced CD8⁺ T cell activity in the tumour (Fong et al., 2020). PBF-509 (taminadenant, NIR178), a 2,6-di (pyrazol-1-yl)pyrimidin-4-amine (Figure 1, compound 20) was well tolerated in non-small cell lung cancer (NSCLC) patients, with manageable adverse effects of immune stimulation, and resulted in tumour shrinkage (either with or without anti-PD-1 monoclonal antibody spartalizumab) (Chiappori et al., 2018). Mixed A_2A/A_2B receptor antagonist AB928 (Figure 1, compound 21) is in clinical trials for cancer immunotherapy, as A_2B receptor activation also has an immuno-suppressive effect in the tumor microenvironment (Seitz et al., 2019). Other A_2A receptor antagonists are in development for PD (e.g., KW6356) or oncology (e.g., EOS-850, ClinicalTrials.gov Identifier: NCT03873883), but their structures are undisclosed (Chen & Cunha, 2020; Vigano et al., 2019).

What are the side effects of the more potent and selective A_2A receptor antagonists? Potential side effects related to the physiological function of A_2A receptors, including sleep disturbance and pro-inflammatory and vasoconstrictor effects, were not seen in early clinical trials of istradefylline (Chen & Cunha, 2020). Clinical testing of an A_2A receptor antagonist, tozadenant (SYN115; Figure 1, compound 13), for Parkinson’s disease was discontinued due to deaths from drug-induced agranulocytosis (Chen & Cunha, 2020), but it is unclear if this a mechanism- or compound-related toxicity. Also, disruption of bone structure was noted in joints of A_2A receptor^−/− mice, and A_2A receptor activation was found to reduce osteoclast formation (Mediero, Perez-Aso, & Cronstein, 2013). Thus, chronic administration of an A_2A receptor antagonist might lead to side effects on joint health. In general, because of the boosting of the immune system by A_2A receptor antagonists, in chronic inflammatory conditions, regular caffeine use might worsen the condition. For this reason, there are proposed applications of selective adenosine receptor agonists for various inflammatory or ischaemic disease states, such as sickle cell disease (Jacobson et al., 2019).

5 |. STRUCTURE-BASED DESIGN OF A_2A RECEPTOR ANTAGONISTS

Prior to the last decade’s advances in protein structure determination, rational design of A_2A receptor ligands was limited to using ligand- or pharmacophore-based approaches or homology models based on rhodopsin, the first published GPCR structure (Palczewski et al., 2000). A major breakthrough was the determination of a β₂-adrenoceptor structure in 2007 (Cherezov et al., 2007; Rosenbaum et al., 2007) and shortly thereafter a structure of the A_2A receptor bound to a prototypical selective antagonist ZM241385, a triazolo {2,3-a}{1,3,5}triazine, (PDB code: 3EML; Figure 1, compound 10) was solved (Jaakola et al., 2008). Comparisons of the A_2A receptor crystal structure to homology models of the complex with ZM241385 revealed that earlier mutagenesis and modelling efforts had correctly identified the orthosteric ligand binding site, but the ligand binding mode was difficult to predict with high accuracy (Michino et al., 2009). This demonstrated that access to GPCR crystal structures provided valuable information on the mechanisms of ligand recognition.

To date, at least 38 crystal and cryo-EM structures of A_2A receptors are available (Carpenter & Lebon, 2017), which revealed the structural basis of agonism and antagonism. More recently, A₁ receptor X-ray and cryo-EM structures were reported (Cheng et al., 2017; Draper-Joyce et al., 2018). As in earlier modeling predictions (Ivanov et al., 2009), adenosine primarily interacts with A_2A receptor residues in transmembrane helices (TM) 3, 6, 7, and the second extracellular loop (ECL2). Compared to the inactive receptor structures, agonist binding leads to rearrangements in TM3 and TM5 in the binding site, which triggers conformational changes in the cytoplasmic region that enable G protein coupling. The adenine moiety of adenosine forms hydrogen bonds with Asn253^6.55 and Glu169^ECL2, whereas the ribose interacts with Ser277^7.42 and His278^7.43 (using conventional residue numbering, Ballesteros & Weinstein, 1995) (Figure 2b). A majority of the A_2A receptor crystal structures that have been elucidated are complexes with antagonists (Zhang, Stevens, & Xu, 2015), underscoring the widely held view that crystallization of receptors may be more feasible for antagonist complexes generally. These illustrate the value of studying ligand interactions at atomic resolution and the variability of antagonist recognition that can occur (Figure 2c–k). All antagonists establish hydrogen bonds with Asn253^6.55 but do not form the interactions with both Ser277^7.42 and His278^7.43 that are characteristic of agonists and block conformational changes observed in agonist-bound structures. Crystal structures of adenosine receptor complexes with low affinity antagonists remain especially challenging. Thus, crystal structures of the human A_2A receptor in complex with caffeine (Cheng et al., 2017; Doré et al., 2011) and theophylline (Cheng et al., 2017) have been important both for establishing the detailed molecular interactions of these small molecules with A_2A receptors and, more generally, for establishing methodologies to determine crystal structures of complexes with weaker affinity antagonists. Weaker affinity antagonists such as caffeine and theophylline also have their practical uses, as they are important investigational compounds added during protein purification to improve the stability of the purified receptor and are often later exchanged with higher affinity ligands of interest. The first complex with caffeine was solved in 2011 (PDB code: 3RFM) (Doré et al., 2011), and a recent high-resolution A_2A receptor structure showed that this compound can bind in two distinct (roughly flipped) orientations (Figure 2c, PDB code: 5MZP) (Cheng et al., 2017). Structures of adenosine receptor complexes with high-affinity xanthines, XAC (Figure 1, compound 6; Figure 2d, PDB code: 3REY) and A₁ receptor-selective PSB36 (Figure 2e, PDB code: 5N2R), revealed that these compounds also bind adenosine receptors with different orientations of the core scaffold (Cheng et al., 2017; Doré et al., 2011). The unexpected variation of xanthine binding modes and diversity of antagonist scaffolds (Figure 2f–k) clearly illustrate the complexity of rational drug design. As one of the first therapeutically relevant GPCRs that was crystallized, the A_2A receptor has served as a prototype for exploring structure-based drug design more generally among class A GPCRs.

FIGURE 2.

Crystal structures of the human A_2A receptor in complex with adenosine and antagonists. (a) Receptor structure shown as grey cartoons with the orthosteric binding site highlighted with a transparent blue oval. The antagonist ZM241385 (also shown in Figure 1, as compound 10) is shown as sticks with orange carbon atoms and residue Asn253^6.55 as sticks with white carbon atoms (PDB code: 4EIY, Liu et al., 2012). The bound sodium ion (purple sphere) is coordinated by Asp52^2.50 (shown as sticks). Binding modes are shown for purines: (b) adenosine (PDB code: 2YDO, Lebon et al., 2011), (c) caffeine (Figure 1, compound 1; PDB code: 5MZP, Cheng et al., 2017), (d) XAC (Figure 1, compound 6; PDB code: 3REY, Doré et al., 2011), (e) PSB36 (PDB code: 5N2R, Cheng et al., 2017); and for non-purines: (f) ZM241385 (PDB code: 4EIY), (g) Tozadenant (Figure 1, compound 13; PDB code: 5OLO, Rucktooa et al., 2018), (h) Vipadenant (Figure 1, compound 14; PDB code: 5OLH, Rucktooa et al., 2018), (i) HTL-1071 (PDB code: 6GT3), (j) a triazole-carboximidamide antagonist (PDB code: 5UIG, Sun et al., 2017) and (k) parent drug of LUAA47070, a prodrug in the family of compound 19 in Figure1; PDB code: 5OLV, Rucktooa et al., 2018). The ligands and key residues are shown as sticks. Hydrogen bonds are shown as dashed lines. Water molecules are shown as red spheres

The large number of successful studies involving structure-based design of A_2A receptor antagonists suggests that this approach should be considered to be an essential component of the toolbox for drug design. In fact, one of the best examples of the use of structure-based design to develop an actual clinical candidate is the A_2A receptor antagonist HTL-1071 (Figure 1, compound 18). A structure-based virtual screen of 0.5 million compounds was first carried out using a homology model of the A_2A receptor. Experimental evaluation of 230 putative ligands led to the discovery of 20 ligands with significant activity (Langmead et al., 2012). By using an approach, termed biophysical mapping, that combines computational docking and measurement of ligand affinity to receptors with mutations in the binding site, reliable models of receptor-ligand complexes could be derived to guide hit-to-lead optimization (Congreve et al., 2012; Zhukov et al., 2011). Crystal structures determined for the A_2A receptor complexes in this chemical series confirmed the predicted ligand binding mode, and further optimization led to identification of HTL-1071, which is currently being evaluated in clinical trials as an anticancer drug (Jazayeri et al., 2017).

Atomic resolution structures of the A_2A receptor enabled virtual screens for ligands in large compound libraries. In two concurrent studies, Katritch et al. and Carlsson et al. computationally docked commercial chemical libraries with several million compounds to a crystal structure of antagonist-bound A_2A receptors with the goal of identifying novel ligand scaffolds (Carlsson et al., 2010; Katritch et al., 2010). From the first screen, Katritch et al. experimentally evaluated 56 top-ranked compounds and found 23 with affinities better than 10 μM. Carlsson et al. predicted 20 out of 1.4 million docked compounds would bind A_2A receptors and experimentally confirmed seven compounds with promising affinities. The high docking hit rates represented a great improvement over hit rates observed in empirical screening campaigns, and several novel chemical scaffolds with nanomolar affinities were among the discovered ligands. The identified antagonists were predicted to occupy the same binding site as ZM241385 and formed interactions with conserved residues Asn253^6.55, Glu169^ECL2, and Phe168^ECL2. Recent virtual screens of chemical libraries have identified A_2A receptor antagonists with unique selectivity profiles to avoid side effects (Ballante et al., 2020) and specific multi-target activity to achieve improved drug efficacy (Jaiteh et al., 2018).

The ability of A_2A receptors to bind low MW compounds such as caffeine suggested fragment-based lead discovery approaches could prove promising. Screens with libraries containing fragment-sized compounds generally yield higher success rates, but the discovered hits have inherently low affinities. Thus, sensitive screening methods are required, and hits need to be subsequently optimized for improved activity and affinity. To enable fragment-based screening methods with A_2A receptors, the receptor thermostability was increased through systematic amino acid replacement, which permitted screening experiments to be carried out at near ambient temperatures and over sufficiently long periods of time 2011. Screens of fragment libraries against the A_2A receptors using NMR- and Surface Plasmon Resonance (SPR)-based approaches identified both orthosteric and allosteric modulators with hit rates of ~3–10% (Chen et al., 2012; Congreve et al., 2011). Computational methods have also been successful in identifying fragment ligands. Chen et al. docked a commercial fragment library with 0.3 million fragments to the A_2A receptor binding site. Of the 22 experimentally tested fragments, 14 showed significant binding, corresponding to an excellent hit rate of 64% (Chen, Ranganathan, IJzerman, Siegal, & Carlsson, 2013). Encouragingly, there was little overlap between scaffolds identified by biophysical and virtual fragment screening, indicating that the two approaches are complementary.

Molecular dynamics (MD) simulations have the potential to provide more mechanistic representations of receptor-ligand complexes than X-ray structures alone, including conformational changes and interactions with water and membrane (Hollingsworth & Dror, 2018). One application of MD simulations in ligand design is to characterize hydration networks in the receptor binding site. In lead optimization, analysis of hydration sites can guide design of ligand substituents that lead to improved affinity by displacing ordered waters (Bortolato, Tehan, Bodnarchuk, Essex, & Mason, 2013; Higgs, Beuming, & Sherman, 2010). The more rigorous, but also computationally demanding free energy perturbation (FEP) approach was used by Lenselink et al. (2016) to demonstrate that simulations could be used to predict relative binding affinities of A_2A receptor antagonists. This study also included design of a new antagonist with improved nanomolar affinity. Similarly, Matricon et al. (2017) found that FEP calculations can guide hit optimization and identify binding modes of fragment ligands. MD simulations were also used to explore the kinetics of ligand binding to the A_2A receptor. Studies of the recognition of antagonists and their dissociation from the A_2A receptor orthosteric binding pocket have provided clues towards the relationship between the chemical structure of ligands and their pharmacokinetic properties. Molecular dynamics simulations of antagonist dissociation from the A_2A receptor orthosteric pocket proposed key roles for several extracellular-facing residues that modulated ligand off-rates (Guo et al., 2016). Residues Glu169^ECL2 and His264^ECL3 in particular were proposed to form a salt bridge that decreased the off rates of some antagonists (Guo et al., 2016). Subsequent studies of crystal structures of A_2A receptors in complex with ZM241385-related heterocyclic antagonists showed that some complexes exhibited distinct relative side chain orientations of Glu169^ECL2 and His264^ECL3, suggesting a structural basis for variations among ligand dissociation rates (Segala et al., 2016).

The A_2A receptor is a typical GPCR for the development of new biophysical methodologies to study drug-receptor interactions, which can provide mechanistic information on the roles of ligand binding in signal transduction. Among recent developments, NMR spectroscopy has emerged as a powerful tool to investigate the function-related conformational dynamics of GPCRs (Lee, Nivedha, Tate, & Vaidehi, 2019) and the regulation of their dynamic behaviour by low MW compounds. Indeed, NMR spectroscopy has been a major source of information on the structural plasticity of GPCRs (Shimada, Ueda, Kofuku, Eddy, & Wüthrich, 2019). NMR studies have provided experimental data that support the view that GPCRs inherently exist in multiple, simultaneously populated conformations, the relative populations of which are modulated by the binding of drugs. NMR studies using probes of dynamics placed at the A_2A receptor intracellular signalling surface (Eddy, Gao, et al., 2018; Sušac, Eddy, Didenko, Stevens, & Wüthrich, 2018; Ye, Van Eps, Zimmer, Ernst, & Prosser, 2016) have documented multiple conformations of A_2A receptors for both antagonist and agonist-bound complexes and have quantified rates of exchange among different conformers (Sušac et al., 2018). Initial clues into the mechanisms of allosteric regulation of A_2A receptor dynamic behaviour by cations (Ye et al., 2018) and lipids (Staus, Wingler, Pichugin, Prosser, & Lefkowitz, 2019) have come from recent NMR studies, as well as mechanistic information on the ligand-driven activation pathways of GPCRs using stable isotopes distributed throughout the receptor structure (Eddy, Lee, et al., 2018) and methyl NMR probes (Clark et al., 2017).

New biophysical approaches for drug discovery have also been vetted in applications to identify novel A_2A receptor antagonists. NMR screening methods have identified novel antagonist compounds in studies of the native, full-length A_2A receptor utilizing specialized detergents that improved the stability the native receptor (Ignonet et al., 2018) as well as thermostabilized A_2A receptor variants fixed to flow devices (Ignonet et al., 2018). Higher throughput MS-based methods of ligand screening have recently been developed and used to identify novel A_2A receptor antagonists (Lu et al., 2019).

Structural and biophysical studies of A_2A receptor interactions with Na⁺ ions established a structural basis for the widely documented role of these ions as GPCR allosteric modulators (Katritch et al., 2014), which was observed in 1.7 Å (PDB code: 5IU4, Segala et al., 2016) and 1.8 Å (PDB code: 4EIY, Liu et al., 2012) crystal structures of A_2A receptors in complex with the antagonist ZM241385 (Figure 2a). This study revealed that Na⁺ ions formed coordinative bonds with several highly conserved residues including Asp52^2.50, and replacement of these residues with different amino acids resulted in loss of signal transduction without significantly perturbing ligand binding. Further studies of the underlying structural basis for these observations were provided by a crystal structure of an A_2A receptor-D52N variant that presented an altered conformation proximate to the NPxxY motif (White et al., 2018). NMR studies of the A_2A receptor-D52N variant demonstrated that neutralization of the charged residue D52, and the consequential displacement or loss of bound Na⁺ ions, was correlated with arrested or altered dynamics at the intracellular signalling surface (Eddy, Lee, et al., 2018).

6 |. CONCLUSION AND FUTURE OUTLOOK

The psychostimulant caffeine, itself, has beneficial effects when consumed in moderation, as well as undesired health effects, especially at higher doses. Caffeine consumption of <400 mg·day⁻¹ by adults, except in pregnancy, appears to be safe and may provide some therapeutic benefit. Clinical studies of caffeine, both meta-analysis and some prospective studies, suggest health benefits in the aging brain, infectious states, pain, diabetes, and liver and kidney disorders. However, in certain chronic conditions, such as bone degeneration, chronic exposure to caffeine might worsen the condition. The long trail of caffeine research has led to the discovery of its antagonism of the effects of adenosine, largely through the A_2A receptors. Adenosine receptors mediate many but not all of caffeine’s effects, so the clinical effects of more potent and selective adenosine receptor antagonists may or may not be comparable. Furthermore, the effects of the panadenosine receptor antagonists, such as caffeine, are not identical to those of more selective A_2A receptor antagonists. There are now numerous synthetic A_2A receptor antagonists, and their design is now structure-based. X-ray crystallography, cryo-EM, NMR, and computational modelling have all been used to gain exquisite insight into ligand binding and activation of the A_2A receptor. Clinical trials in several disease areas are in progress with naturally occurring alkylxanthines and with synthetic A_2A receptor antagonists. The focus of A_2A receptor antagonist development is currently for oncological and neurological indications, including PD, Alzheimer’s disease, ADHD, and the immunotherapy of cancer. Thus, a fascination with the stimulant effects of caffeine over five decades has led to the development of new therapeutics and a deeper understanding of the A_2A receptor structure as a prototypical GPCR.

6.1 |. Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Harding et al., 2018), and are permanently archived in the Concise Guide to PHARMACOLOGY 2019/20 (Alexander, Christopoulos, et al., 2019; Alexander, Fabbro, et al., 2019; Alexander, Kelly, et al., 2019; Alexander, Mathie, et al., 2019).

Abbreviations:

ADHD

attention-deficit hyperactivity disorder

COPD

chronic obstructive pulmonary disease

cryo-EM

electron cryo-microscopy

ECL

extracellular loop

FEP

free energy perturbation

HDAC

histone deacetylase

KO

knockout

MD

molecular dynamics

NSCLC

non-small cell lung cancer

PD

Parkinson’s disease

SNP

single nucleotide polymorphism

TM

transmembrane helix

TME

tumour micro-environment