A_2A Adenosine Receptor Antagonists in Neurodegenerative Diseases

Abstract

Background:

Alzheimer’s disease (AD) is the most common form of dementia worldwide, with approximately 6 million cases reported in America in 2020. The clinical signs of AD include cognitive dysfunction, apathy, anxiety and neuropsychiatric signs, and pathogenetic mechanisms that involve amyloid peptide-β extracellular accumulation and tau hyperphosphorylation. Unfortunately, current drugs to treat AD can provide only symptomatic relief but are not disease-modifying molecules able to revert AD progression. The endogenous modulator adenosine, through A_2A receptor activation, plays a role in synaptic loss and neuroinflammation, which are crucial for cognitive impairment and memory damage.

Objective:

In this review, recent advances covering A_2A adenosine receptor antagonists will be extensively reviewed, providing a basis for the rational design of future A_2A inhibitors.

Methods:

Herein, the literature on A_2A adenosine receptors and their role in synaptic plasticity and neuroinflammation, as well as the effects of A_2A antagonism in animal models of AD and in humans, are reviewed. Furthermore, current chemical and structure-based strategies are presented.

Results:

Caffeine, the most widely consumed natural product stimulant and an A_2A antagonist, improves human memory. Similarly, synthetic A_2A receptor antagonists, as described in this review, may provide a means to fight AD.

Conclusion:

This review highlights the clinical potential of A_2A adenosine receptor antagonists as a novel approach to treat patients with AD.

1. INTRODUCTION

Alzheimer’s disease (AD), an age-associated pathology, is one of the principal causes of dementia in the elderly and the fifth main cause of death in patients in the ≥65 year age group. The number of people in America suffering from AD reached an impressive 5.8 million in 2020 and is expected to triple by 2050 [1]. Two types of AD have been described that include sporadic and familial forms, presenting late and early onset, respectively, with the latter being responsible for less than 1% of cases [1]. In the case of sporadic AD, the principal risk factor is age, followed by an ApoE-ε4 genetic polymorphism.

In familial AD, mutations affecting amyloid precursor protein (APP), and presenilin 1 and 2 (PS1 and PS2) genes, provoke beta-amyloid peptide (Aβ) accumulation and plaque formation with a more rapid pathological development [2]. The clinical signs of AD include cognitive dysfunction, apathy, anxiety and neuropsychiatric signs [3, 4]. The pathogenetic mechanisms of AD involve Aβ extracellular accumulation and hyperphosphorylation of tau protein [5]. Aβ peptides originate from amyloid precursor protein (APP) following sequential cleavage by β- and γ-secretases [2, 6]. According to the amyloid hypothesis, the Aβ plaque generation is the main cause of AD, inducing neurotoxicity and synaptic dysfunction, and it is responsible for memory decline [7]. During the pathological progression, Aβ aggregation increases phosphorylation of tau, the most important microtubule-related protein, playing a crucial role in microtubule formation and being relevant for neuronal plasticity and axonal outgrowth [8, 9]. Hyperphosphorylated tau protein is released from microtubules and produces intracellular, neurotoxic, and insoluble neurofibrillary tangles (NFT-s) [10]. All of these processes induce neuronal death and neuroinflammation, the major hallmarks of AD [11]. Approved drugs for AD primarily increase acetylcholine (ACh) transmission as well as reduce glutamate excitotoxicity, and comprise donepezil, rivastigmine, galantamine and memantine [12, 6]. However, these drugs can induce beneficial effects in the short term but are not able to stop or reverse AD progression. Furthermore, these drugs have various liabilities upon long-term use. The production of effective disease-modifymg drugs able to impede or cure AD appears to be a difficult goal to reach, as demonstrated by the numerous molecules that entered but eventually failed in clinical trials [4, 13]. Recently, a novel antibody decreasing extra-neuronal Aβ protein accumulation in the brain, called aducanumab, has been approved by the US Food and Drug Administration (FDA) [14]. However, the utility of this therapy is under debate due to the conflicting results of the two clinical trials, EMERGE and ENGAGE, carried out to test its efficacy in patients with mild cognitive impairment (MCI) and early dementia. There is considerable skepticism from the scientific community regarding this new and expensive Biogen drug, aducanumab [15–17]. Overall, the management of AD has become a global concern due to a worldwide increase in life expectancy [1]. Unfortunately, AD incidence, prevalence and mortality, as well as the costs of patient care, and its influence on society, are dramatically increasing [1]. Therefore, in the context of exploring new therapies to treat AD, in this review, we will discuss the potential involvement of the A_2A adenosine receptor in this pathology, leading to the suggestion that A_2A antagonists might be a strategic approach to discover new drugs.

2. A_2A ADENOSINE RECEPTORS IN AD

A_2A receptors belong to the P1 purinergic receptor family of four G protein-coupled receptors (GPCRs) that respond to the endogenous modulator adenosine, comprising A₁, A_2B and A₃ subtypes [18]. One structural characteristic of the A_2A subtype not shared by other receptor family members is a long intracellular carboxy terminal tail, where phosphorylation and palmitoylation processes may induce receptor desensitization and internalization [19]. This subtype is able to associate with different receptors, such as A₁ adenosine and D₂ dopamine receptors, to form heteromers having distinct pharmacological characteristics with respect to monomers [20]. The A_2A receptor is distributed in both the central nervous system and the periphery. The highest expression is in the striatum, the olfactory tubercle, the immune system, and at lower levels in the cerebral cortex, hippocampus, heart, lung, and vasculature [21].

At the neuronal presynaptic level, the main effect of A_2A receptor activation is to increase glutamate release, thus contributing to excitotoxicity. Postsynaptically, the A_2A receptor promotes alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionate (AMPA) activation and N-methyl-D-aspartate (NMDA) phosphorylation mGluR5-dependent (Fig. 1). In astrocytes, the A_2A receptor provokes activation, proliferation and reduction of glutamate uptake, by regulating the expression of glutamate transporters GLT-1 and GLAST. In microglia, its activation is responsible for proliferation and the release of inflammatory mediators [22]. In the peripheral immune system, the A_2A receptor is present at high levels in almost all immune cells, e.g., neutrophils, monocytes, macrophages, dendritic and T cells, as well as platelets, and blood vessels, where it produces important non-redundant anti-inflammatory, antiaggregatory, and vasodilatory actions, respectively [23]. A_2A receptor stimulation triggers activation of G_olf proteins in the striatum and G_s, proteins in both the brain and the periphery. Thus, it is responsible for increasing cyclic AMP and PKA phosphorylation, as well as regulation of downstream Akt and MAPK pathways [24–29]. Due to the widespread distribution of the A_2A receptor and the increase of its endogenous ligand in both inflammation and cancer, it affects diverse pathologies spanning neurodegenerative, autoimmune, inflammatory and malignant diseases. Specifically, adenosine controls multiple processes in the brain, from sleep to seizures, and cognitive and memory functions, through A₁ and A_2A receptor activation to modulate the activity of excitatory glutamatergic synapses [30].

Fig. (1).

Schematic overview of A_2a receptor signaling involved in brain-related mechanisms of AD pathology.

Under basal activity, adenosine activates the A₁ receptor, inducing a reduction in hippocampal glutamate signaling. However, it may also stimulate presynaptic A_2A receptors, which remove the A₁-mediated inhibition, thus increasing glutamate release [31]. At the postsynaptic level, the A_2A receptor facilitates long-term potentiation (LTP) by regulating NMDA receptor activation and intracellular Ca²⁺ [22, 32–35]. It is well known that synaptic dysfunction and degeneration in the temporal lobe represents the first hallmark of cognitive disability, prior to Aβ plaque and tangle formation [36]. Indeed, patients with MCI and early AD present a loss of synapses in the hippocampus and in the posterior cingulate gyrus, suggesting that this event represents an initial point leading to memory damage [37–39].

Due to the involvement of the A_2A adenosine receptor in glutamatergic synaptic physiology, a link between this subtype and AD has been revealed It has been reported that in aging, AD animal models and AD patients, A_2A adenosine receptor expression is increased in both hippocampal neurons and astrocytes [30, 40–46], thus raising glutamate release [47, 48], calcium influx, long-term potentiation (LTP)-to- long-term depression (LTD) shift [49], and cognitive impairment [50]. Of historical interest, the first indication that the A_2A receptor might be increased in the Alzheimer’s brain was evident m a comparison of specific A_2A receptor photoaffinity labeling in post--mortem striatal membranes [51]. In contrast, A_2A receptor antagonism attenuates hippocampus-dependent memory disabilities and LTP alterations in aged animals [52, 53] and AD models [54–57]. In addition, genetic silencing of A_2A adenosine receptors can ameliorate synaptic damage present in AD models [58–60]. Interestingly, adenosine levels in postmortem AD brains are greater in the parietal and temporal lobes compared to the frontal cortex, suggesting increased A_2A receptor stimulation in overexpressing regions [61]. The level of A_2A receptor expression in human correlates with disease states, and thus, it is considered a biomarker reflective of susceptibility and progression of brain diseases [62]. The correlation of single nucleotide polymorphisms (SNPs) of the A_2A receptor with neuropsychiatric and neurodegenerative disorders has been documented [62].

Another important A_2A adenosine receptor function involved in AD concerns its modulation of neuroinflammation through its effects on glial cells [63, 64]. In particular, astrocytes of AD patients displayed elevated levels of the A_2A adenosine receptor, and its genetic knockdown in young and aging mice increased long-term memory [44]. Functionally, A_2A receptors in astrocytes were essential for the fine-tuning of inhibitory and excitatory modulation of synaptic transmission by affecting GABA and glutamate uptake. Specifically, A₁ and A_2A receptors, present as heteromers, regulated GABA transport in an opposite fashion, with the A₁ inhibiting and the A_2A promoting it [65]. In addition, the A_2A receptor, which is essential for the reduction of glutamate transporters GLAST and GLT-I induced by Aβ peptide, impaired glutamate uptake [66]. A_2A receptor upregulation in primary cortical astrocytes was recently reported to alter the astrocytic transcriptome with an important effect on genes relevant for inflammation and angiogenesis [67]. In addition to astrocytes, cells relevant to neuroinflammation that overexpress the A_2A receptor are activated microglia, which also express another AD target receptor, i.e., the NMDA receptor. It has been reported that A_2A and NMDA receptor subtypes interact, producing a novel entity, overexpressed in hippocampal cells from the APPSw, Ind mice, characterized by cross-antagonism, where A_2A receptor antagonism may hamper the overstimulation of NMDA receptor activity [68, 69]. In addition, this approach to counter A_2A receptor-mediated cytokine release may attenuate neuroinflammation to ameliorate memory dysfunction [70, 71]. Several Works have reported the relevance of A_2A adenosine receptor antagonists for the restoration of function following the synaptic loss and cognitive disability in animal models of AD, suggesting a strategy to counteract synaptic toxicity [44]. Interestingly, caffeine, the most consumed A_2A receptor antagonist following chronic ingestion through coffee and other foods, as well as genetic removal of A_2A receptors in KO (knockout) mice, reduced hippocampal tau hyperphosphorylation, counteracted neuroinflammation and reverted the related memory deficit [56, 59, 72, 73]. Accordingly, in animal models of AD, exposure to low doses of the selective A_2A antagonist, istradefylline (a commercially available new co-therapy for Parkinson’s disease (PD)), increased spatial memory and habituation, giving an important proof-of-concept that A_2A receptor blockade might be a novel target to fight cognitive impairments in AD patients [74, 75]. This result suggests a great potential for drug development against dementia targeting the A_2A receptor, for example, by repurposing istradefylline. Istradefylline has passed multiple clinical safety studies and is already approved in Japan, Korea and the US (as Nouriast) [76–79].

3. MEDICINAL CHEMISTRY OF A_2A RECEPTOR ANTAGONISTS

The focus on A_2A receptor antagonist therapeutic development was for PD during the 1990s and 2000s, which has now shifted toward immuno-oncology, as well as other neurodegenerative diseases, such as AD [80]. A wide range of chemotypes is now known to antagonize the A_2A receptor [81, 82]. The appeal of A_2A receptor antagonists to drug discovery programs is also strengthened by the precedent of widespread use of caffeine without serious adverse effects, and the non-lethal and well-tolerated phenotype of A_2A receptor-KO mice. The ability of A_2A receptor-selective PET tracers to determine the pharmacodynamics and receptor occupancy in the human brain has also aided drug discovery in this field [83].

The chemical modification of caffeine (1,3,7-trimethylxanthine, 1, Fig. 2) and naturally-occurring alkylxanthines has a long history [73], which predates knowledge of adenosine receptors. Early therapeutic targets included treatment of asthma (by analogy to the very efficacious antiasthmatic drug theophylline (Table 1), which has since fallen out of favor as a preferred treatment) and intermittent claudication (e.g., pentoxifylline). From 1968 to 1980, it became apparent that caffeine at μM concentrations blocks the action of adenosine at its newly discovered receptor(s) [84, 73]. Although a nonselective adenosine receptor antagonist, caffeine, is being examined clinically for the treatment of AD [clinicaltrials.gov, NCT04570085, accessed July 26, 2021], earlier studies indicated that caffeine intake correlates inversely with the occurrence of AD [85, 86] but a recent analysis found no correlation [87]. Currently, the major focus of selective A_2A antagonist development is on immuno-oncology, and some previous candidate molecules for PD treatment have been repurposed for cancer [88]. However, it is conceivable that these anticancer agents, if successful, could later generate renewed interest for neuroprotection.

Fig. (2).

Structures of key A_2A adenosine receptor ligands, including antagonists that are now or were recently in clinical trials for PD and/or immuno-oncology.

Table 1.

Affinity selected A_2A AR antagonists mentioned in the text (K_i values in nM from binding assays, human, unless noted, (r), rat).

-	Ki				-
-	A1AR	A2AAR	A2BAR	A3AR	Ref.
CGS15943	3.5	0.15	71	51	[117, 82]
-	6 (r)	1.2 (r)	-	-
ZM241385	255	0.8	50	>10,000	[82]
SCH58261	594	1.1	>10,000	>10,000	[117]
SCH442416	1111	0.048	>10,000	>10,000	[118]
	1815 (r)	0.50 (r)	-	>10,000 (r)
Preladenant	1474	1.1	>1700	>1000	[108, 119]
XAC	6.82 1.2 (r)	18.4 63 (r)	7.75	25.6 >10,000 (r)	[120]
Istradefylline	2830	36	1800	>3000	[121, 117]
CSC	>10,000	38	8200	>10,000	[117, 122]
Imaradenant	160	1.7	64	>10,000	[123]
Taminadenant	2500	12	1000	5000	[124]
Ciforadenant	192	3.54	1530	2460	[125]
Etrumadenant	64	1.5	2.0	489	[126]
Vipadenant	68	1.3	63	1005	[127]
Tozadenant	1350	5.0	700	1570	[82]

The screening of many xanthine analogues at the rat A₁ and A_2A receptors became feasible with the advent of adenosine receptor-specific, high-affinity radioligands. Although moderately selective for the rat A₁ receptor [89], [³H]xanthine amine congener (XAC, Table 1) was of sufficient affinity to be demonstrated in human platelets as the first A_2A receptor antagonist radioligand [90]. [³H]NECA was useful as an agonist A_2A radioligand in rat brain membranes when A₁ receptors were blocked selectively [91]. Other radioligands and PET ligands (for in vivo imaging of the brain A_2A receptor) followed [81, 83, 92–95]. Antagonist selectivity for the A₁ receptor was achieved first among the four receptor subtypes [96], followed in 1993 with the first A_2A receptor-selective xanthine antagonists, the 8-styrylxanthines. Substitution of the xanthine scaffold can be directed toward increased affinity at each of the four adenosine receptors [96]. For the A_2A receptor, the 8-styrylxanthines have achieved moderate A_2A selectivity [97], e.g., 8-(3-chlorostyryl)-caffeine (CSC), which also inhibits MAO-B [98]. The most notable 8-styrylxanthine was introduced as KW-6002, now known as istradefylline [97, 76]. Based on its A_2A antagonism and bioavailability in the brain when taken orally, istradefylline is currently approved as a co-therapy for PD with levodopa/carbidopa that reduces off-time episodes in patients [99].

One of the earliest efforts to identify A_2A receptor-selective non-xanthine antagonists led to reports of CGS15943, a slightly A_2A receptor-selective antagonist, of which the [1, 2, 4]triazolo[1,5-c]quinazolin-5-amine scaffold was inspired by multiple nitrogen substitutions of the xanthine scaffold and a phenyl ring fusion [100]. Screening of chemical libraries, first by experimentally-based or high throughput screeningand eventually by structure-based in silico methods, greatly increased the diversity of chemotypes known to bind to the A_2A receptor selectively [101–103].

As mentioned above, CGS15943 showed high A_2A receptor affinity but only marginal receptor subtype selectivity, and its initial chemical modification was not successful in achieving high selectivity. In fact, N⁵-acyl derivatives of CGS15943 have been shown to have high affinity at the human A₃ receptor, and to some degree, the human A_2B receptor. An early parallel effort by the pharmaceutical industry identified ZM241385 as a moderately selective A_2A antagonist bearing a [1, 2, 4]triazolo[2,3-a] [1, 3, 5]triazin-5-ylamino] scaffold [104]. ZM241385 was later shown to be potent at the A_2B receptor [105]; thus, ZM241385 is a less than optimal pharmacological tool compound for the A_2A receptor. However, it has been utilized in the first and numerous subsequent A_2A receptor X-ray structures, as described below [106, 107]. Its p-hydroxy-(2-phenylethyl) amino moiety points toward the extracellular regions in the human A_2A receptor structures, although its precise position in this flexible region of the A_2A receptor protein has proven variable.

In addition to CGS15943 and ZM241385, a heterocyclic scaffold that proved highly productive in the search for novel A_2A antagonists was the 7H-pyrazolo[4,3-e] [1, 2, 4]triazolo[1,5-c]pyrimidin-5-amines, first introduced by Baraldi and colleagues, an academic lab research group working closely with the pharmaceutical industry [81]. Among the early antagonists bearing this scaffold were SCH58261 and SCH442416 [26, 92, 108]. Further derivatization of this scaffold identified the congener Preladenant (SCH420814), which progressed through Phase 3 clinical trials for PD [109]. Unfortunately, only a subset of the clinical results indicated efficacy, so its intended use in PD was abandoned. Later, the same compound was repurposed for cancer immunotherapy as MK-3814, as a monotherapy or in combination with an anti-PD-1 monoclonal antibody (Pembrolizumab) [88].

The elucidation of the three-dimensional structure of the human A_2A receptor, initially in 2008 and followed by dozens of X-ray and cryo-electron microscopic structures, has greatly facilitated the discovery of A_2A receptor antagonists [106, 110, 111]. The X-ray structure of the receptor complex with caffeine was reported in 2017 [112]. In fact, the human A_2A receptor has become one of the most highly probed GPCR structures, after the β₂ adrenergic receptor, and the result of this detailed knowledge of its structure and function at the atomistic level (including at a resolution of 1.8 Å that reveals specific water molecules in the binding site [107]) has enabled the identification of many new potential therapeutic agents [73, 111]. The X-ray structures and free energy calculations can guide the modification of known A_2A antagonists to enhance affinity in a rational, structure-based fashion [113, 114].

A_2A antagonist AZD4635 (HTL1071, Imaradenant), a 5,6-diaryl-1,2,4-triazin-3-amine derivative, is in clinical trials for cancer immunotherapy [115]. Prior to its application to cancer treatment, it was considered for the treatment of ADHD. It was the first A_2A antagonist designed by structure-based methods that has progressed to human testing, displaying a K_i value of 1.7 nM human A_2A receptor and showing >30-fold selectivity over other adenosine receptors [111].

Other scaffolds that provide A_2A antagonists, including some that have entered clinical trials, include (Fig. 2) Vipadenant (BIIB014, a [1–3]triazolo[4,5-d]pyrimidin-5-amine); Tozadenant (SYN115, a 4,7-disubstituted benzo[d]thiazole); Taminadenant (NIR178, PBF-509, a 2,6-disubstituted pyrimidin-4-amine); Ciforadenant (CPI-444, V81444, a [1–3]triazolo[4,5-d]pyrimidin-5-amine); EOS100850 (structure not disclosed); and mixed A_2A/A_2B antagonist Etrumadenant (AB928, a 1,2,3-triazol-4-yl-pyrimidine). Tozadenant caused unanticipated toxicity in a clinical trial for PD, leading to five deaths from drug-induced agranulocytosis out of 409 patients [79]. Similar toxicity had not been observed with other A_2A antagonists in clinical trials. An A_2A antagonist of undisclosed structure (Inupadenant, EOS-850) is in Phase 1 clinical trial for use against solid tumors [116].

CONCLUSION

The A_2A adenosine receptor is one of the major players coming from the purine field in the treatment of neurodegenerative diseases, such as PD and AD. It shows various pharmacological properties, including antioxidant, anti-inflammatory, cytoprotective, antitumoral, antiplatelet, hepatoprotective, and antifibrotic activities. Due to its relevant biological effects, numerous chemical design approaches have been reported. Diverse heterocycles have been reported as potent and selective A_2A antagonists, and some have entered clinical trials, either for PD or immuno-oncology. Nonselective antagonist caffeine is being examined as a clinically useful agent for AD. Despite the clinical trials of various selective A_2A antagonists for PD, only one with FDA approval is a caffeine analogue, istradefylline. However, the clinical use of this antagonist offers the opportunity for further exploration of its neuroprotective properties. From another point of view, due to the complex nature of neurodegenerative diseases, it is obvious that “simple” targeting of A_2A adenosine receptor antagonists for their therapy would not be enough. Thus, a multitarget-directed ligand approach involving A_2A adenosine receptor antagonists has been suggested. However, the rational design of compounds that interact with several targets is very challenging. Up to now, dual-target A_2A antagonists/D₂ agonists, A_2A /MAO-B inhibitors as well as triple-target A₁/ A_2A/MAO-B blockers and A₁/A_2A/H3 antagonists have been reported [122, 128–131]. This review offers a comprehensive overview of A_2A antagonist receptor synthesis in the context of drug discovery for AD therapy.