|Home | About | Journals | Submit | Contact Us | Français|
Pharmacological tool compounds are now available to define action at the adenosine (ARs), P2Y and P2X receptors. We present a selection of the most commonly used agents to study purines in the nervous system. Some of these compounds, including A1 and A3 AR agonists, P2Y1R and P2Y12R antagonists, and P2X3, P2X4 and P2X7 antagonists, are potentially of clinical use in treatment of disorders of the nervous system, such as chronic pain, neurodegeneration and brain injury. Agonists of the A2AAR and P2Y2R are already used clinically, P2Y12R antagonists are widely used antithrombotics and an antagonist of the A2AAR is approved in Japan for treating Parkinson’s disease. The selectivity defined for some of the previously introduced compounds has been revised with updated pharmacological characterization, for example, various AR agonists and antagonists were deemed A1AR or A3AR selective based on human data, but species differences indicated a reduction in selectivity ratios in other species. Also, many of the P2R ligands still lack bioavailability due to charged groups or hydrolytic (either enzymatic or chemical) instability. X-ray crystallographic structures of AR and P2YRs have shifted the mode of ligand discovery to structure-based approaches rather than previous empirical approaches. The X-ray structures can be utilized either for in silico screening of chemically diverse libraries for the discovery of novel ligands or for enhancement of the properties of known ligands by chemical modification. Although X-ray structures of the zebrafish P2X4R have been reported, there is scant structural information about ligand recognition in these trimeric ion channels. In summary, there are definitive, selective agonists and antagonists for all of the ARs and some of the P2YRs; while the pharmacochemistry of P2XRs is still in nascent stages. The therapeutic potential of selectively modulating these receptors is continuing to gain interest in such fields as cancer, inflammation, pain, diabetes, ischemic protection and many other conditions. Reported potencies refer to the human receptors unless otherwise noted. Additional affinity data can be found in recent review papers (Müller and Jacobson, 2011; Jacobson et al., 2015; Coddou et al., 2011a).
The effects of extracellular purines and pyrimidines at their receptors in the central and peripheral nervous systems have been under intense research scrutiny. Tools that can be used in this effort, in addition to genetic knock-out or knock-down of receptor expression include a vast collection of directly acting agonists and antagonists, allosteric modulators of the receptors, and indirect modulators that affect the level of endogenous agonists present. This review will concentrate on selective agonists and antagonists of the adenosine receptors (ARs), P2Y receptors (P2YRs) and P2X receptors (P2XRs), and in particular compounds that are readily available to the research community. The principle endogenous agonists are adenosine for the ARs and ATP for the P2XRs, while at the P2YRs a variety of adenine and uracil nuclotides have been shown to be native activators. These native P2YR agonists include ATP, ADP, UTP, UDP, UDP-sugars and some dinucleoside polyphosphates. Some compounds that might be even more selective than the ones discussed here might be under development, but they are not treated in the present work in detail. Reported potencies refer to the human receptors unless otherwise noted.
Numerous selective agonists of the four subtypes of ARs (A1, A2A, A2B and A3 ARs, Table 1) and their precursors have been used in studies of the nervous system (Chen et al., 2013), and a selection of the many ligand analogues, both directly acting agonists (2 – 21), antagonists (31 – 63), and indirect modulators (26 – 31), is presented here (Figures 1 and and2).2). Adenosine itself 1 is a native, nonselective AR agonist that is short-lived in the body; while its metabolite inosine 2, following the action of adenosine deaminase, weakly activates the A3AR (Gao et al., 2011). A hybrid molecule, abbreviated NECI, resembling both inosine and the potent nonselective agonist NECA (adenosine 5′-N-ethyluronamide) was shown to have enhanced affinity at the A3AR (van Galen et al., 1994). Extracellular adenosine is produced indirectly from ATP 4 via AMP 3 by the sequential action of ectonucleotidases. Thus, release of ATP and other nucleotides under stress conditions generally results in increased AR activation. Nucleotides such as ATP 4 are generally inactive at ARs, although AR agonist effects have been ascribed to AMP 3, either as an intact molecule or as a ready precursor for locally produced adenosine through the action of ecto-5′-nucleotidase (Bhattarai et al., 2015). Other studies indicate that direct AR activation by AMP itself would not occur at sub-μM concentrations (van Galen et al., 1994). Regadenoson (CVT-3146), a selective agonist of the A2AAR used to induce stress in cardiac imaging, and istradefylline (KW-6002), a xanthine antagonist of the A2AAR is approved in Japan for treating Parkinson’s disease.
AR agonists, particularly those selective for A1AR or A3AR, have shown neuroprotective effects in stroke and other models (Zylka et al., 2011; Fishman et al., 2012; Rivera-Oliver and Díaz-Ríos, 2014). However, the precursor nucleotide molecules such as ATP can be damaging in neuroprotective models by activating P2Rs, and it is more commonly observed that P2Y or P2X antagonists are more protective than P2 agonists. The native agonist adenosine 1 is short lived in vivo, and therefore it has only limited use in models of neuroprotection. However, most of the synthetic analogues of adenosine shown in Figure 1 are more stable biologically and suitable for in vivo administration. All of the AR knockout mice have been generated and none are lethal.
AR subtype selective agonists are generally synthetic adenosine derivatives that are long lasting metabolically.
A1AR agonists tend to be substituted at the N6 position with arylakyl (5), cycloalkyl (6 – 9), bicycloalkyl (10), or aryl (11, 12) substituents (Müller and Jacobson, 2011). The C2 and 5′ positions also may be substituted with Cl, 7 and 9, respectively, or the 5′ position with cyclic moieties (12). Compound 5 was one of the first AR agonists to be widely used. It was previously defined as a moderately A1AR selective agonist, but its use in that capacity is discouraged in favor of more selective nucleosides. A1AR agonists 6a and 6b are widely used pharmacological probes with 6b being more selective for the A1AR. A cautionary note is that many of the selective A1AR agonists used routinely such as 6a and 6b have considerable activity at the A3AR. Compound 6b and agents that increased endogenous adenosine mimicked the acute antinociceptive effect of acupuncture, consistent with the hypothesis that the A1AR mediates this effect (Goldman et al., 2010). The agonist INO-8875 7 (also known as Trabodenoson), which is an A1 agonist in Phase III clinical trials for glaucoma, is also intended for optic neuropathy.
A1-selective nucleoside derivative 8 failed to show efficacy in a clinical trial for dental pain and was tested in a Phase II trial for postherpetic neuralgia or peripheral nerve injury before being discontinued (Elzein and Zablocki, 2008). A1-selective nonnucleoside derivative capadenoson (BAY68-4986) 9 is in trials for treatment of persistent atrial fibrillation and is now available as a research tool (Tendera et al., 2012). Compound 10 is highly selective for the A1AR and displayed analgesic activity in the formalin test in mice (Luongo et al., 2012, Franchetti et al., 2009). Compound 11 is a peripherally-selective agonist, due to its permanently charged sulfonate group, that is moderately A1AR selective (70-fold). Compound 12 was tested in a clinical trial for peripheral neuropathic pain but discontinued (Ochoa-Cortes et al., 2014). Compound 13 displayed antinociceptive effects in mice and was suggested to cross the blood brain barrier and act directly through the A1AR despite the 5′-phosphate group (Korboukh et al., 2012). Partial agonists of the A1AR (such as the 2′-deoxy analogue of 6) are effective in chronic neuropathic but not acute pain models and lack some of the cardiovascular side effects of full A1AR agonists (Schaddelee et al., 2005). Compound 14 is a moderately selective full A1AR agonist that protects in some seizure models (Tosh et al., 2012b). It is well tolerated in rodents and does not produce the toxicity upon dose escalation that is typical of more widely used A1AR agonists, likely from cardiovascular effects. Compound 14 was shown to induce anti-depressant effects through an increase of homer1a in the brain (Serchov et al., 2015).
A2AAR agonists are often substituted at the C2 position with arylalkyl (15) or ethynyl (16) substituents. Compound 15 is a widely used pharmacological probe for activation of the A2AAR, but its high AR subtype selectivity seen in rat and mouse is reduced at human (h) ARs, and its degree of entry into the brain is low. A single intrathecal injection of 15 or 16 was reported to have a long duration of protection against mechanical allodynia and thermal hyperalgesia in a model of chronic constriction injury (CCI) in mice (Loram et al., 2009). An X-ray structure of the hA2AAR complex with agonist 15 was recently reported (Lebon et al., 2015). Another A2AAR agonist BVT.115959 (structure not disclosed) advanced to clinical trials for pain (Ochoa-Cortes et al., 2014). Compound 17 is a selective A2AAR agonist that was in clinical trials for COPD, which were discontinued due to lack of efficacy. With its large molecular weight and many H-bonding groups, it is not orally bioavailable. However, these molecular features increase the stability of the A2AAR complex, such that it was possible to obtain an X-ray crystallographic structure without extensive stabilizing point mutations (Xu et al., 2011). This was the first X-ray structure of an AR complex with an agonist to be reported. Although not detected in the X-ray structures, the A2AAR can form in situ various functional homo- and hetero- (e.g. with the A1AR or the D2 dopamine receptor) di- or multimers (Ferré et al, 2014; Bonaventura et al, 2015; Navarro et al., 2015). The A1AR can also heterodimerize with the P2Y1R.
Compound 18 has been used as a nonnucleoside agonist for selective activation of the A2BAR, but its potency, selectivity and efficacy are less than originally reported (Wessam et al., 2015; Hinz et al, 2014). Thus, it should be used cautiously and in combination with appropriate antagonists. 18 is of the same chemical series (3,5-dicyanopyridines) as selective A1AR agonist 9.
A3AR agonists (19 – 24) are generally adenosine derivatives substituted at the C2, N6 and 5′ positions with various substituents. Those groups most favorable for A3AR selectivity include: N6 - benzyl (19, 20, 22 – 24) or small alkyl (21); C2 - H, Cl, alkynyl or arylethynyl. Compounds 19 (also known as Piclidenoson) and 20 are widely used pharmacological probes with 20 being more selective for the A3AR, and they are in clinical trials for inflammation (rheumatoid arthritis and psoriasis) and primary liver cancer, respectively (Fishman et al., 2012). Compound 21 has been used as a highly selective A3AR radioligand of high affinity (Klotz et al., 2007). The actions of A3AR agonists 19 and 24 in chronic neuropathic pain and cerebroprotection have been described (Little et al., 2015). They completely reverse or prevent allodynia and hyperalgesia in the chronic constriction injury (CCI) model but have no effect on acute pain. Compounds 22 – 24 contain a conformationally constrained bicyclic ring in place of ribose with a North conformation (N)-methanocarba ring system, which adds to the A3AR selectivity (Tosh et al., 2012a, 2014, 2015). Compound 23 is a peripherally-selective agonist that is highly A3AR selective (Paoletta et al., 2013). The pharmacokinetics of A3AR agonist 24, administered intraperitoneally in rat, indicates a t1/2 of 1.09 h (Tosh et al., 2015), consistent with its duration of action in vivo in the CCI pain model. 24 is also orally active in the CCI model.
Compounds that indirectly modulate the action or levels of extracellular adenosine include: inhibitors of adenosine deaminase (e.g. Pentostatin, 25); inhibitors of adenosine kinase (e.g. nucleoside 5-iodotubercidin, 26 and nonnucleoside ABT-702, 27); and positive allosteric modulators (PAMs, e.g. A1AR-selective T-26, 28 and TRR469, 29) (Müller et al., 2012). Compound 27 decreased both chronic and acute pain and acted when administered either peripherally or centrally (Kowaluk et al., 2000). It advanced toward clinical trials but the process was discontinued. Compound 29 suppressed pain in a manner comparable to and additive with morphine in formalin and writhing tests and had an antiallodynic effect in the streptozotocin-induced model of diabetic neuropathic pain (Vincenzi et al., 2014). The structure of 29 was derived from the prototypical enhancer 2-amino-3-benzoylthiophene PD-81,723 (structure not shown). The combination of A1AR enhancer 29 and agonist 6b inhibited the release of excitatory amino acid analogue D-aspartate in spinal cord synaptosomes. A positive allosteric modulator of the hA3AR 30 was shown to enhance the actions of both adenosine and inosine at this subtype (Gao et al., 2011).
AR antagonists are shown in Figure 2. Various AR antagonists, in particular those selective for A2AAR, have shown neuroprotective effects in models of Parkinson’s and Alzheimer’s disease and some other models (Navarro et al., 2015; Rivera-Oliver M and Díaz-Ríos, 2014; Armentero et al., 2011; Chen, 2014). Selective AR antagonists have been used as tool compounds in studies of neurodegeneration and pain. In general, the affinity of selective AR antagonists at the other AR subtypes and species differences should be taken into consideration when selecting a dose (Müller and Jacobson, 2011; Alnouri et al., 2015).
The widely ingested alkylxanthines theophylline 31 (1,3-dimethylxanthine) and caffeine 32 (1,3,7-trimethylxanthine) at physiological concentrations are weak (affinity ~10 μM) nonselective antagonists the ARs, except that the affinity is particularly low at rodent A3AR. The main caffeine metabolite paraxanthine 33 (1,7-dimethylxanthine) also acts as a central nervous stimulant and with even higher potency than caffeine at ARs and might be useful for treating hypersomnia associated with neurodegenerative diseases (Okuro et al., 2010). 8-p-Sulfophenyl derivatives 34 and 35 are moderately potent nonselective antagonists of ARs with the added benefits of water solubility and restricted entry to the brain.
A1AR-selective antagonists are derived from nonxanthines, e.g. adenine derivative 36 and pyrrolopyrimidine 37. An 8-cycloalkylxanthine derivative 38 is the most widely used A1AR-selective antagonist with nM affinity, but its A1AR selectivity is greater in rodents than in human. Several 8-bicycloalkylxanthine derivatives 39 and 40 are even more selective for the A1AR.
An 8-styrylxanthine derivative 41 is approved for Parkinson’s disease treatment in Japan. The related compound 42 is selective for the A2AAR, but also inhibits monoamino oxidase-2 (MAO-2), two activities that would be beneficial in neurodegenerative diseases. Two A2AAR-selective antagonists, 48 and 51, are commonly used as A2AAR-selective pharmacological probes of nM affinity. The X-ray structure of A2AAR complex with antagonist 48 was the first antagonist-bound AR structure to be reported (Jaakola et al., 2008), and now a higher resolution structure of 1.8 Å resolution is available (Liu et al., 2012). In the same family as 51 is compound 50, which was in clinical trials for treatment of Parkinson’s disease. A2AAR-selective antagonists 46 and 47 were also in clinical trials for the same condition. The A2AAR affinity of 51 was initially reported in the subnanomolar range, but the Ki value was redetermined to be 4 nM (Müller and Jacobson, 2011). The A2AAR-selective adenine derivative 52, with a relatively low molecular weight, is effective in Parkinson’s disease models. A2AAR-selective antagonists are also of interest in the context of checkpoint blockade for cancer immunotherapy, possibly by coadministration with other anticancer agents (Leone et al., 2015).
Compounds 53–58 are A2BAR-selective antagonists (Müller and Jacobson, 2011), and in general the affinity among these xanthine derivatives is greater at the human A2BAR than at the rodent A2BAR. The first A2BAR-selective antagonists to be reported were 53 and 54. Subsequently, other 1,3-dialkylxanthines with 8-aryl substitution 55–58, including the highly water-soluble derivative 57, were found to be A2BAR-selective (Borrmann et al., 2009) Compounds 53 and 54 were introduced by the pharmaceutical industry for possible application to asthma, but neither remains on a clinical path.
An effective and moderately selective A3AR antagonist for use in mouse and rat is 62 (Li et al., 1998; Kreckler et al., 2006). Various other heterocyclic antagonists of the A3AR, such as MRS1191 59a, MRS1334 59b, MRS1220 60, MRS3008-F20 61, PSB-10 63a and PSB-11 63b, in most cases have much higher affinity at the human than the murine A3ARs and should be used cautiously in pharmacological experiments utilizing rodents.
Although radioligand binding assays are most commonly used in the discovery of new ligands, the use of fluorescent AR agonists and antagonists is increasing (Kozma et al., 2013; Stoddart et al., 2015). Other specialized tool compounds for ARs include irreversibly binding agonists and antagonists (Jacobson et al., 1989; Shryock et al., 1998), which have been used to detect receptor reserve.
The only one of the ARs to be probed extensively using X-ray crystallography is the A2AAR. There are currently 14 crystal structures of the human A2AAR, including four structures of the complex with the selective antagonist ZM241385 48 (Xu et al., 2011; Liu et al., 2012; Lebon et al., 2015; Congreve et al., 2015). The receptor has been crystallized with a various agonists and both high and low affinity antagonists using receptor constructs that are stabilized by mutagenesis at multiple single sites. The X-ray crystal structures are useful for structure-based studies at ARs leading to the design of new ligands and for the modeling of closely related GPCRs, i.e. the other three ARs (Paoletta et al., 2013; Tosh et al., 2012a). For example, two of the A2AAR antagonist-bound crystal structures are results of the application of the receptor structure for the screening and identification of novel triazine-based ligands, one of which has entered a clinical path for treating attention deficit/hyperactivity disorder (Congreve et al., 2015).
The eight subtypes (P2Y1, P2Y2, P2Y4, P2Y6, P2Y11, P2Y12, P2Y13, and P2Y14) of P2YRs are gaining attention in studies of the nervous system (Weisman et al., 2012; Puchałowicz et al., 2014; Brunschweiger et al., 2006). All of the P2YR knockout mice have been generated (and some double P2YR knockouts) and studied except for the P2Y11R, which appears to be absent in the mouse (Table 2). P2YRs have been found to regulate the adaptation of the central nervous system to ischemia, tissue damage, inflammation and chronic neurodegenerative diseases (Burnstock and Verkhratsky, 2012). Various P2YR ligands were studied for their effects in models of neuropathic pain and migraine (Andó et al., 2010; Magni and Ceruti, 2013). This section will focus on the currently available agonists and antagonist for these receptors. Ligand discovery at the P2YRs has lagged behind the discovery of AR ligands, but recently many agonist and antagonist ligands have been reported. Nevertheless, not all of the P2YR subtypes have selective agonists and antagonists; thus, in order to characterize the P2YR subtype(s) involved in a particular phenomenon, pharmacological studies often require sequential use of various agonists (64 – 85, Figure 3) and antagonists (86 – 109, Figure 4) that may not be definitive for a single P2YR subtype. Some of the compounds are limited in their application due to their enzymatic lability in biological systems and restricted bioavailability. Thus, there remains a need to introduce a wider range of P2YR ligands of diverse structure and pharmacological selectivity. A few P2YR drugs are approved for clinical use. In addition to three P2Y12R antagonists that are widely used antithrombotics, a nonselective agonist of the P2Y2R is used clinically in Japan for treating dry eye disease.
The native agonists include: ATP, 4; ADP, 64; UTP, 72; UDP, 78; UDP-glucose (UDPG), 84. These nucleotides are generally short-lived in the body, and thus the interconversion (including enzymatically mediated formation and deactivation) of P2YR agonists in situ has been a complicating factor from the outset of studies of P2YRs and purine/pyrimidine receptors in general. Various inhibitors of enzymes that process extracellular nucleotides have been reported recently (Bhattarai et al., 2015; Lee et al., 2015; Chang et al, 2014; Al-Rashida and Iqbal, 2014; Baqi, 2015; Corbelini et al., 2015), according to the strategy that the activation of P2Rs (and ARs) can be modulated effectively by indirect means (Zimmermann et al, 2012).
The receptor subtypes that are preferentially activated by 5′-diphosphates are: P2Y1, P2Y12 and P2Y13 (ADP); and P2Y6 (UDP) receptors. The P2Y14R is dually activated by UDP-sugars, such as UDP-glucose (84), and also by UDP 78, although this combination of agonists was not recognized at first as the cloning of this receptor concluded that only UDP-sugars are active (Chambers et al., 2000). Only later, was the ability of UDP to activate the P2Y14R established (Carter et al., 2009a). The receptor sybtypes that are preferentially activated by 5′-triphosphates are: P2Y2 and P2Y4 and P2Y11. ATP 4 acts at all three subtypes, and furthermore UTP 72 activates the P2Y2 and P2Y4Rs. The P2Y4R has a functional difference between species that causes ATP 4 and its derivatives to act as partial agonists (or antagonists) at the human homologue and full agonists at the rat homologue (Herold et al., 2004). Various naturally occurring dinucleotides also active P2YRs, i.e. the asymmetric dinucleoside tetraphosphate Up4A (Jankowski et al., 2005), which is a vasoconstrictor, and a range of symmetric diadenosine polyphosphates ApxA (Jankowski et al., 2003).
The P2Y1R is activated by adenine, but not uracil nucleotides. A 2-methylthioadenosine-5′-diphosphate analogue 68 that has a rigid substitution of the ribose ring in the form of a (N)-methanocarba ring system, has been shown to be a potent and selective agonist of the P2Y1R in comparison to the P2Y12R and P2Y13R (Chhatriwala et al., 2004). This ring maintains a P2Y1R-preferred conformation, which is responsible for the high potency and selectivity. However, the potent 2-methylthio-substituted agonist 65, 2-MeSADP (riboside), which is widely used in pharmacological studies, acts at all three subtypes. Naturally occurring dinucleotides 70 and 71 (dinucleoside tetraphosphates) activate the P2Y1R, but not selectively (Jankowski et al., 2009; Durnin et al., 2014; Yelovitch et al., 2012). For example, Ap4A 70 also activates the P2Y2R (former known as P2U; Lazarowski et al., 1995). The biology of such dinucleotides in vasodilation, vasoconstriction and other cardiovascular activities has been explored. The concentration of Ap4U 71 in human plasma is ~56 nM, which is relevant to activation of P2YRs (Jankowski et al., 2009). Ap4U also activates the P2Y2R (Wiedon et al., 2012).
Antagonists of the P2Y1R have been sought as potential anti-thrombotic agents, although none have been approved for clinical use. P2Y1R antagonists include both nucleotides, e.g. 86 – 88 (Jacobson et al., 2015), and nonnucleotides, e.g. 89, which represents a large group of diarylurea derivatives reported in the past several years. The urea derivative 89, which is commercially available, has a Ki value of 6 nM at the P2Y1R (Chao et al., 2013); it and many of its analogues tend to be hydrophobic in nature. On the other hand, very high affinity has been achieved in the hydrophilic series of nucleotides. The principle by which nucleotide derivatives antagonize the P2Y1R instead of activating it involves formalistically splitting the 5′-diphosphate and attaching a monophosphate group as a phosphoester at the 3′-position, removal of the 2′-hydroxyl group, methylation of the exocyclic amino group and other modifications (Boyer et al., 1996; Jacobson et al., 2015). All of these modifications contribute to the loss of efficacy in this chemical series, and the antagonism of the bisphosphates applies to P2Y1R homologues species-independently. The most potent nucleotide antagonist of the P2Y1R 88, which also contains a potency-enhancing (N)-methanocarba ring system, has subnanomolar affinity. Thus, the (N)-conformation of the ribose-like ring at the P2Y1R is a requirement for both nucleotide agonists and antagonists. A corresponding South conformation (S)-isomer in the bisphosphate series is greatly reduced in affinity (Jacobson et al., 2015).
While, in general, 5′-diphosphates are more potent and efficacious, ATP 4 acts as a partial agonist or agonist, respectively, at the P2Y1R and P2Y12R (Waldo and Harden, 2004; Paoletta et al., 2015). ATP-γ-S 67 is also an agonist at various P2YRs, including the P2Y1R (Waldo and Harden, 2004). The γ-thiophosphate group impedes (Malmsjö et al., 2003), but does not completely prevent its enzymatic hydrolysis. A caged form of ATP 69, which regenerates ATP upon irradiation has been studied and is potentially useful for studies of P2Rs (Amatrudo et al., 2015). Similarly, the 1-(3,4-dimethyloxy-6-nitro-phenyl)ethyl (DMNPE) group was used to block the β-phosphate of agonist 65 to abolish its activity at the P2Y1R and P2Y12R, and this activity could be recovered upon irradiation (Gao et al., 2008). Such caged compounds are potentially useful for studying rapid neuronal responses to P2Y (or P2X) receptor activation.
P2Y1R is present on both astrocytes and neurons in the brain. P2Y1R activation produces anxiolytic effects in the rat elevated plus-maze (Kittner et al., 2004). The P2Y1R is a possible target for the treatment of traumatic brain injury (Choo et al., 2013). An astrocytic P2Y1R was found to be hyperactivated in an Alzheimer’s disease model in the mouse (Delekate et al., 2014).
Both UTP 72 and UTP-γ-S 74 activate the P2Y2R and P2Y4R, and the γ-thiophosphate group of 74 increases stability toward ectonucleotidases. However, the γ-thiophosphate group of 74 and 67 introduces chemical lability toward oxidation when left exposed to the air. Compound 75 is a potent and selective P2Y2R agonist that contains a triphosphate mimic consisting of a β,γ-dihalomethylene bridge. δ-Blocked 5′-tetraphosphate derivative 76 is a selective P2Y2R agonist with moderate (μM) potency. Few effective P2Y2R antagonists and none of nM affinity are readily available. A complex derivative of 4-thiouracil 91 is a P2Y2R antagonist with μM potency (Kemp et al., 2004). There are no selective P2Y4R antagonists, but a number of P2Y4R agonists were reported, of which 77 is representative. There is a tolerance for steric bulk at P2Y2R, P2Y4R and P2Y6R in the receptor-bound uracil nucleotides that have an alkyloximino group at the 4-position of the nucleobase (Jacobson et al., 2015). Thus, they are more properly cytosine derivatives; 5′-CTP itself is only weakly active at the rat P2Y2R and antagonizes the hP2Y4R (Kennedy et al., 2000). The precise selectivity in this chemical series of 5′-di- and triphosphates among these three P2YR subtypes can be largely modulated by structural changes of the 4-alkyloximino group (Jacobson et al., 2015). Nonselective dinucleotide P2Y2R/P2Y4R agonist diquafosol (INS365, Up4U) was approved in Japan (2010) for dry eye disease. Another nonselective dinucleotide P2Y2R/P2Y4R agonist denufosol (INS37217, Up4dC) has been evaluated in phase III trials for bronchial indications, but was not approved.
Numerous derivatives of UDP, e.g. 78 – 81, have shown selectivity as agonists of the P2Y6R. Certain dinucleotides (dinucleoside triphosphates), e.g. 82, potently activate the P2Y6R (Shaver et al., 2005). This dinucleotide also contains a 4-alkoxyimino group on one of the two pyrimidine nucleobases and is highly potent, but only moderately selective. Only one compound 92 is currently used as a selective antagonist of the P2Y6R. Unfortunately, it is poorly soluble in water and of limited chemical stability due to its two isothiocyanate groups, which are essential for its antagonist activity (Mamedova et al., 2004). UDP activates the P2Y6R in microglial cells to induce phagocytosis (Koizumi et al., 2007). A novel P2Y6R agonist protects glial cells against apoptosis (Haas et al., 2015). High affinity P2Y6R fluorescent agonists MRS4129 (selective) and MRS4162 (also potent at P2Y2R, P2Y4R) are useful in flow cytometry (Jayasekara et al., 2014).
Both ATP and ATP-γ-S 74 activate the P2Y11R, with the latter being more potent, but not selective. Recently, β,γ-dichloromethylene analogues of a 2-substituted derivative of ATP were reported to be agonists of the P2Y11R (Haas et al., 2013). A selective nonnucleotide agonist NF546 83 of the P2Y11R is available (Meis et al., 2011). Curiously, this derivative of the nonselective P2R antagonist suramin 96 activates, rather than antagonizes the P2Y11R. Other derivatives of highly charged suramin antagonize the P2Y11R: 97 and 98.
A huge number of nucleotide and nonnucleotide antagonists of the P2Y12R have been reported (Paoletta et al., 2015). The reason for this focus on P2Y12R antagonists is their value as antithrombotic agents. They include: β,γ-dihalomethylene bridged 5′-triphosphate mimics (93, 94); uncharged nucleotide like derivative 95 that is now used clinically (ticagrelor); sulfonated anthraquinone derivatives, e.g. 99; uncharged heterocycles 104 (previously in clinical trials as an antithrombotic) and 108 (commercially available). Ticagrelor 98 is the first competitive P2Y12R antagonist approved as an antithrombotic drug and was the product of a long program in pharma industry to circumvent the need for anionic groups in such antagonists (Springthorpe et al., 2007; Hoffman et al., 2014). The X-ray structure of the P2Y12R complexes with nonnucleotide antagonist 108 and with nucleotide agonist 65 and partial agonist 66 were recently reported (Zhang et al., 2014a, 2014b). These structures showed dramatic changes in conformation of the extracellular regions, in comparing forms with nonnucleotide or nucleotides bound. The nucleotides-bound structures are greatly contracted in the upper regions surrounding the ligand, with the highly positively charged extracellular loops forming electrostatic coordination with the phosphate groups. Thienopyridine P2Y12R antagonists clopidogrel 100 and prasugrel 102 are widely used antithrombotic agents, which must be preactivated by enzymatic transformation in the liver to active metabolites, 101 and 103, respectively. The active metabolites have limited chemical stability, but compound 103 is available commercially. Radiolabeled nucleotide agonist 65 is a suitable tracer for binding studies in cells overexpressing the P2Y12R, and it can also serve as an agonist radioligand for the P2Y1R (Zhang et al., 2014a, 2015). Radiolabeled 93, known as [3H]PSB-0413, is a selective nucleotide antagonist radioligand for the P2Y12R (Ohlmann et al., 2013).
ADP activates the P2Y12R in microglial cells to induce chemotaxis (Haynes et al., 2006). Thus, it serves as a “find me” signal, while UDP acts as an “eat me” signal. P2Y12R antagonists are a target for control of inflammatory and chronic neuropathic pain (Horváth et al., 2014).
Highly selective P2Y13R agonists and antagonists are unknown, but various nucleotide antagonists of the P2Y12R, e.g. 93 and 94, also activate the P2Y13R. At the P2Y13R, one compound 107 is currently used as a selective antagonist. This diazo derivative of pyridoxal phosphate is closely related to the disulfonate derivatives 105 and 106, which are useful antagonists at various P2YRs and P2XRs.
Nonnucleotide antagonists of the P2Y14R have been reported (Robichaud et al., 2011). One naphthoic acid derivative 109 has been useful in pharmacological studies, although this chemical series was rejected for development by the pharmaceutical industry due to a lack of oral bioavailability. Compound 109 displays high selectivity for the P2Y14R in comparison to all other P2YRs (Barrett et al., 2013). Because of its low absorption, attempts were made to design prodrugs of 109 for in vivo use (Robichaud et al., 2011). Other analogues of compound 109 were reported to have potent antagonist activity at the P2Y14R, including MRS4174, a fluorescent conjugate of AlexaFluor488 (structure not shown, Ki 0.8 nM). MRS4174 is a useful tracer for flow cytometric characterization of binding to the P2Y14R (Kiselev et al., 2014). The SAR of nucleotide derivatives at the P2Y14R has been extensively explored, including the report of a high affinity P2Y14R fluorescent agonist MRS4183 (structure not shown, Kiselev et al., 2015). This fluorescent agonist can be prepared from available starting materials in a single reaction step.
The X-ray structure of the P2Y1R complexes with antagonists 88 and 89 were recently reported (Zhang et al., 2015). Unexpectedly, both of these ligands bound at sites distal from the conventional binding region for small molecules deep in the TM helical domain. Nucleotide 88 bound to a more exofacial site in close contact with the extracellular loops, and nonnucleotide 89 bound on the outer surface of the receptor in contact with the phospholipid bilayer. The locations of the urea derivative caused a reexamination of the assumption that it was a competitive inhibitor. It proved to be allosteric with respect to agonist 65, which was used as a radioligand in dissociation kinetic experiments in conjunction with site-directed mutagenesis. The binding sites on the P2Y1R of 88 and 89 were completely distinguishable by mutagenesis studies, consistent with the X-ray structures. Another apparently allosteric P2Y1R inhibitor is 2,2′-pyridylisatogen tosylate (PIT, 90, Gao et al., 2004), although this was not show using mutant P2Y1Rs, and the site for its binding is undetermined.
Docking of nucleotides in a homology model of the P2Y14R, based on the X-ray structure of the P2Y12R in complex with nucleotide agonist 65, was consistent with the observed SAR (Trujillo et al., 2015). Also, both forms of X-ray structures of the P2Y12R have been helpful in characterizing the specific protein-ligand interactions for a wide variety of P2Y12R agonists and antagonists (Paoletta et al, 2015). This suggested that the new P2YR X-ray structures are suitable templates for structure-based design of new ligands. The possibility of hetero- and homodimerization of P2YRs has not yet been addressed structurally with X-ray crystallography, but there is pharmacological support for this phenomenon, e.g. in P2Y1R/P2Y11R heterodimers (Ecke et al., 2008).
P2XRs are ion channels permeable for Na+, K+ and Ca2+ which are activated by ATP (Coddou et al., 2011a). They show a wide distribution in the body. A variety of orthosteric and allosteric ligands for P2X receptors have been reported, some of which are highly charged molecules and therefore not orally bioavailable, but chemically diverse classes of drug-like P2X antagonists are under development (Gunosewoyo and Kassiou, 2010). The P2XR family in mammals consists of seven different subunits, P2X1–P2X7, that form homo- or heterotrimeric channels (Hausmann et al., 2015). There are currently no ligands that are highly selective for different combinations of P2X subunits. Homomeric P2X5R and P2X6R do not to seem functional in humans. P2X2/3R heteromers represent a well established heteromeric P2XR subtype expressed in the dorsal root ganglia. P2XRs are involved in nerve transmission, pain sensation and inflammation (Coddou et al., 2011a; Di Virgilio, 2015; Bele & Fabbretti, 2015; Burnstock, 2015). Prolonged receptor activation can lead to pore formation, and the cells become permeable for large molecules (molecular mass up to ca. 1000 Da). This has been especially observed for the P2X7R (Di Virgilio, 2007). That receptor subtype is also exceptional insofar as it requires high, millimolar ATP concentrations to be activated, while the other subtypes are typically activated at low micromolar or high nanomolar concentrations of ATP (see Table 3).
Two X-ray structures of the zebrafish P2X4R with high resolution were published, one in the closed state, and one with ATP bound (Kawate et al., 2009; Hattori et al., 2012; Grimes & Young, 2015) The previously suggested trimeric structure was confirmed (Nicke et al., 1998). The receptor has a chalice-like structure, and each of the three subunits contains a large rigid ectodomain, two transmembrane domains and intracellular C- and N-termini. There are three equivalent extracellular ATP-binding sites, in about 40 A distance from the cell membrane, located at the subunit interfaces of the trimeric receptor. ATP binding induces the opening of the ion channel by bending of the lower part of the receptor thereby expanding the region near the ion channel pore (Hattori et al., 2012). Mutagenesis studies were useful for the identification of key residues important for receptor-ligand interaction, receptor modulation and ion channel function (Hausmann et al., 2015). Subsequent molecular modeling studies have been performed based on experimental data (Dal Ben et al., 2015). It would still be of paramount interest (i) to obtain X-ray structures of the native human P2X4R, and (ii) to identify binding sites of allosteric modulators by determination of X-ray co-crystal structures.
P2XR function can be allosterically modulated by ions (e.g. Mg2+, Ca2+, Zn2+) (Coddou et al., 2011b; Müller, 2015), steroids (De Roo et al., 2003) and lipids (Bernier et al., 2013), e.g., phosphatidylinositol polyphosphates (e.g., PI(4,5)P2) (Bernier et al., 2008a,b; Mo et al., 2009) The PIPs appear to bind to positively charged amino acids on the cytosolic C-terminal domain. The P2X4R and to some extent also the P2X2R are positively modulated by high concentrations of ethanol (ca. 100 mM) (Yi et al., 2009; Ostrovskaya et al., 2011).
The binding site of the physiological agonist ATP - the so-called orthosteric binding site - is well conserved in the different P2XR subtypes. It contains five positively charged amino acid residues (one Arg and four Lys) which interact with the triphosphate chain (Chataigneau et al., 2013). All potent agonists known so far that bind to the ATP binding site are negatively charged. Also, many orthosteric antagonists are highly polar compounds which cannot penetrate into the CNS. Therefore, and also because of the fact that numerous ATP binding proteins are found in the body, the development of positive and negative allosteric modulators (PAMs and NAMs) appears to be more promising.
Most of the known P2XR agonists are structurally derived from the physiological agonist ATP (Lambertucci et al., 2015). Agonists with high selectivity for a single subtype are presently not available. ATP (4) displays potency in the low micromolar range at all subtypes, except for the P2X7R, which requires millimolar ATP concentrations for activation. 2-Methylthio-ATP (66) and γ-thio-ATP (67) display a similar profile as ATP, but both compounds are metabolically more stable. α,β-Methylene-ATP (111, Figure 5) shows a preference for P2X1 and P2X3Rs with somewhat lower potency at P2X4 and much lower potency at P2X7R, while β,γ-Methylene-ATP (112) is most potent at P2X1R and shows only negligible potency at the other subtypes. Benzoyl-ATP (BzATP, 110) is more potent as compared to ATP with highest potencies at P2X1 and P2X3R subtypes. It should be kept in mind that the activity of allosteric modulators may be probe-dependent, i.e. their potency is likely to depend on the employed agonist (Müller et al., 2012). Therefore, the physiological agonist ATP should preferably be used when characterizing positive or negative allosteric modulators.
Ivermectin (113, Figure 5) a macrocyclic lactone used in veterinary medicine as an antiparasitic agent binds with high affinity to glutamate-gated chloride channels in nerve and muscle cells of the parasites and prevents their closure (Omura & Crump, 2004). It is lipophilic, accumulates in cell membranes and penetrates into the CNS. Ivermectin interacts with many different ion channels (Zemkova et al., 2014). It acts as a PAM at P2X4R facilitating the opening and retarding the closing of the channel. Effective concentration are about 100 nM - 3 μM (Lalo et al., 2007). The exact binding site and mechanism of allosteric enhancement of the P2X4R remain speculative (Müller, 2015). Recently, it was found that ivermectin is also a PAM at the hP2X7R at similar concentrations (3 μM), but not at rat and mouse P2X7Rs (Nörenberg et al., 2012).
Within several series of antagonists some compounds with positive modulatory activity (PAMs) were discovered. This shows that small modifications can turn an allosteric inhibitor or NAM into a PAM.
MRS2219 (114) was found to selectively potentiate ATP-induced responses at recombinant rat P2X1Rs expressed in Xenopus oocytes with an EC50 value of 5.9 μM (Jacobson et al., 1998). Compound 114 was inactive at rat P2X2, P2X3 and P2X4Rs.
Among a series of P2X2 antagonists with an anthraquinone core structure several derivatives were identified that showed positive allosteric modulation. PSB-10129 (115) was one of the most potent PAMs leading to a 3-fold maximal increase in the ATP-elicited current with an EC50 value of 489 nM at the hP2X2R (Baqi et al., 2011).
The anthraquinone derivative Cibacron Blue (116. Figure 6) acts as a PAM of hP2X3Rs. It showed a 3–7-fold increase in the magnitude and the potency of ATP activating Ca2+ influx and transmembrane currents with an EC50 value of 1.4 μM (Alexander et al., 1999). Compound 116 was also found to be a PAM at the rat P2X4R at concentrations of 3–30 μM, while it blocked the receptor at high concentrations of 300 μM (Miller et al., 1998). Thus, the compound is non-selective and may also interact with other P2XR subtypes.
The development of PAMs may be a promising approach for the development of drugs to achieve an increase in P2XR activation.
Moderately potent, non-selective P2XR antagonists include suramin (96), Reactive Blue 2 (RB-2, 117), PPADS (105), and iso-PPADS (106). These compounds are of limited use and should be replaced by more potent and selective antagonists that are now available. The ATP derivative TNP-ATP (118) is very potent at P2X1 and P2X3R (low nanomolar IC50 values), and somewhat less potent at P2X2 and P2X4R. It is virtually inactive at P2X7R.
Effects, especially those related to CNS activity, are outlined below for P2X1, P2X2, P2X3, P2X4 and P2X7Rs. The effects of P2X5 and P2X6Rs in the nervous system are mostly unexplored; there are no selective ligands for those subtypes.
P2X1Rs are expressed in smooth muscle cells of various organs including arteries, and vas deferens (Mulryan et al., 2000). P2X1 knockout mice were resistant to thromboembolism suggesting the P2X1R to be a potential target for thrombosis and stroke prevention (Hechler & Gachet, 2011). The P2X1R is also expressed in neuronal and glial cells, and a strong upregulation was found after mechanical CNS injury. Inhibition of P2X1Rs resulted in neuroprotective effects after ischemic or toxic CNS injury. Thus, the P2X1R was suggested as a drug target for neuroprotective action (Hausmann et al., 2012). P2X1 antagonists have also been suggested for the treatment of Parkinson’s disease since they might reduce alpha-synuclein accumulation (Navarro et al., 2015).
Only few P2X1 antagonists are known so far (Figure 6). These include suramin and derivatives: NF279 (119) and NF449 (120). Compound 120 is significantly more potent than 119 with subnanomolar potency at P2X1 and very high selectivity. It is therefore the preferred P2X1 antagonist. They appear to be competitive antagonists (Hausmann et al., 2012; El-Ajouz et al., 2012; Soto et al., 1999; Rettinger et al., 2000; Kassack et al., 2004). Their interaction with the receptor has recently been studied by a mutagenesis approach (Farmer et al., 2015).
RO0437626 (RO-1, previously also known as RO116-6446, 121) is a non-acidic P2X1 antagonist, with moderate potency (IC50 recombinant hP2X1, calcium assay, 3 μM) (Jaime-Figueroa et al., 2005; Ford et al., 2006). It shows selectivity versus hP2X2, hP2X3 and hP2X2/3Rs (IC50 > 100 μM). In a rat model it attenuated bladder contractions at 1 or 10 μmol/kg i.v. (King et al., 2004).
P2X2Rs show a wide distribution in the peripheral and the central nervous system, as well as on many non-neuronal cell. They play a role in sensory transmission and the modulation of synaptic function. P2X2R are slowly desensitized and are the only homomeric P2XR subtype that is potentiated by acids. Potentiation is also observed by Zn2+, but other divalent cations inhibit the receptor at high concentrations (Gever et al., 2006). P2X2 subunits form homotrimeric or heterotrimeric channels with P2X3Rs. P2X2R knockout mice displayed a reduced pain response (Cockayne et al., 2005). Potential indications for P2X2 antagonists include pain, and they may also show neuroprotective properties.
Only few selective antagonists for P2X2R have been developed. The anionic standard P2R antagonists PPADS (105), RB-2 (117), TNP-ATP (118) and suramin (96) are moderately potent, non-selective P2X2R antagonists. The nucleotide 118 was shown to act as a competitive antagonist while suramin binds to an allosteric site as shown in radioligand binding studies at rat P2X2Rs (Trujillo et al., 2006). The suramin derivative NF770 (122, Figure 6) is more potent and selective for P2X2Rs, and evidence was presented that it exhibits a competitive mechanism of action (Wolf et al., 2011). Potent and selective P2X2 antagonists related to the anthraquinone derivative Reactive Blue 2 (117) were developed, PSB-10211 (123) and PSB-1011 (124), which was characterized as a competitive P2X2R inhibitor, although an allosteric mode of action could not completely be excluded (Baqi et al., 2011).
P2X3 subunits may form homomeric or heteromeric (P2X2/3)Rs. They are expressed on nerve cells in the central peripheral nervous system. Particularly high expression is found on small nociceptive sensory and sympathetic neurons. Knock-out mice showed, among other effects, reduced temperature sensation and pain responses (Coddou et al., 2011). P2X3R expression in dorsal root ganglia is upregulated after sciatic nerve ligation in the CCI model. P2X3 knock-out mice are less sensitive to pain stimuli (Jarvis et al., 2003). P2XR antagonists have therefore potential for the treatment of various kinds of pain including neuropathic pain and migraine. They may also be useful for the treatment of epilepsy and sleep disorders.
Many potent, selective and drug-like P2X3R antagonists were developed. Since 2001 more than 50 patents have been filed which claim small molecules as P2X3 and/or P2X2/3R antagonists (Bölcskei & Farkas, 2014; Ford, 2012; Müller, 2010). However, most of the compounds have only been published in the patent literature and only few research reports have appeared so far.
One of the first described selective P2X3 antagonists was A-317491 (125, Figure 6), a tricarboxylate, which binds to the ATP site acting as a competitive antagonist (Jarvis et al., 2002). The compound displays low peroral and CNS bioavailability and high plasma-protein binding but may still be useful as a pharmacological tool compound (Sharp et al., 2006).
A series of allosteric antagonists was derived from the drug trimethoprim discovered in a high-throughput screening campaign by Roche (see Müller, 2010). RO-3 (126), RO-4 (127) and RO-51 (128) belong to that series. RO-4 (later named AF-353, 127) displays IC50 values of 6 nM (hP2X3), 13 nM (rat P2X3) and 25 nM (hP2X2/3), and is selective versus P2X1, P2X2, P2X5, P2X7 (IC50 > 10 μM) as well as a wide range of other receptors and enzymes. Compound 127 shows peroral bioavailability and penetrates into the brain (Carter et al., 2009b). The compound was obtained in 3H-labeled form and used to label its allosteric binding site (Gever et al., 2010). Compound 127 was further modified resulting in the more potent and polar AF-906 (RO-51, 128), which showed superior pharmacokinetic properties (Jahangir et al., 2009). AF-906 (128) displayed IC50 values of 2 nM (hP2X3) and 5 nM (hP2X2/3). Clinical trials have been performed with another member of the series, AF-219 (structure not disclosed), for chronic, treatment-refractory cough, and the outcome was positive (Ford et al., 2013).
The physiological heptapeptide spinorphin (LVVYPWT, 129), which inhibits the effect of encephalin-degrading enzymes resulting in analgesic activity, was reported to act as a potent allosteric antagonist at P2X3Rs. It showed an IC50 value of 8.3 pM in patch clamp studies at the recombinant hP2X3R expressed in Xenopus oocytes, and was virtually inactive at the mouse P2X1 and the hP2X7Rs (Jung et al., 2007).
P2X4Rs are widely expressed in the central nervous system and in the periphery, e.g., in microglia, and on endothelial cells. Peripheral nerve injury leads to microglial activation in the spinal cord which results in increased P2X4R levels (Tsuda et al., 2003; Beggs et al., 2012). This process leads to neuropathic pain. P2X4R knockout mice displayed reduced pain and no development of allodynia. P2X4R antagonists are therefore developed for the treatment of neuropathic pain. Further potential indications include spinal cord injury (de Rivero Vaccari et al., 2012), epilepsy (Ulmann et al., 2013), stroke, multiple sclerosis, and neurodegenerative diseases such as Parkinson’s and Alzheimer’s (Varma et al., 2009).
The serotonin reuptake inhibitor paroxetine (130, Figure 7) was found to act as an allosteric antagonist of P2X4Rs at high concentrations with IC50 values at the rat and the hP2X4Rs of 2.45 μM (rat), and 1.87 μM (human), respectively (Nagata et al., 2009; Müller, 2010). Paroxetine was subsequently investigated in a rat model of neuropathic pain. Intrathecal administration showed an anti-allodynic effect when applied 7 or 14 days after spinal nerve injury. In contrast, the antidepressant amitriptyline which is clinically used for treating neuropathic pain blocked P2X4Rs only weakly (Sim et al., 2010).
The benzodiazepine derivative 5-BDBD (131) is a moderately potent (IC50 0.5 μM), selective allosteric P2X4 antagonist (Donnelly-Roberts et al., 2008; Balázs et al., 2013). The compound displays a very low water-solubility and is therefore not easy to handle.
The allosteric P2X4 antagonist N-(benzyloxycarbonyl)phenoxazine (PSB-12054, 132) exhibited an IC50 of 0.189 μM at the hP2X4R, but was less potent at the rat (2.10 μM) and the mouse P2X4Rs (1.77 μM) (Hernandez-Olmos et al., 2012). It showed >50-fold selectivity for the human P2X4 versus hP2X2, hP2X3 and hP2X7Rs, and >30-fold selectivity versus the hP2X1R. PSB-12054 is currently one of the most potent antagonists at hP2X4Rs. The carbamate structure of PSB-12054 was found to be hydrolytically stable. A drawback is its high lipophilicity and moderate water-solubility.
A more water-soluble analogue is PSB-12062 (133), the N-(p-methylphenylsulfonyl)-substituted phenoxazine. It was about similarly potent at human (IC50 1.38 μM), rat (0.928 μM) and mouse P2X4Rs (1.76 μM) and showed selectivity versus P2X1, P2X3 and P2X7Rs (Hernandez-Olmos et al., 2012)
A series of carbamazepine derivatives was studied as P2X4 antagonists (Tian et al., 2014). The most potent derivative was N,N-diisopropylcarbamazepine (134) displaying an IC50 value at the hP2X4R of 3.44 μM. However, the compound was weaker at mouse and rat P2X4Rs and was not very selective versus hP2X1 and hP2X3Rs, whereas it was very selective versus hP2X2 and hP2X7Rs.
Recently, a urea derivative, BX-430 (135) identified by screening of a compound library as an allosteric antagonist with an IC50 value of 0.54 μM at the hP2X4R and selectivity versus the other P2XR subtypes. The compound had no effect on mouse and rat P2X4Rs (Ase et al., 2015).
Several patents have been filed presenting further P2X4 antagonists, e.g. by Nippon Chemphar and Kyushu University, Japan.
The P2X7R is highly expressed on immune cells, such as macrophages, mast cells, and microglial cells, and can also be found on oligodendrocytes. The receptor is a potential drug target for the treatment of inflammation including neuroinflammatory diseases, pain (neuropathic, inflammatory, nociceptive and chronic), multiple sclerosis, neurodegenerative disorders, cerebral ischemia, brain and spinal cord injury, depression, anxiety and bipolar disorders (Bartlett et al., 2014; Bartlett & Sluyter, 2014; Gunosewoyo et al., 2009, Navarro et al., 2015; Chrovian et al., 2014). P2X7 ligands may also be effective in the treatment of cancer, including brain cancers (Burnstock & Di Virgilio, 2013). The P2X7 is only activated by high, millimolar ATP concentrations, indicating that it has a high relevance under pathological conditions, and pore formation is typical upon its prolonged activation
In recent years much effort has been invested in the development of selective P2X7R antagonists, which have been reviewed recently (Mehta et al., 2014; Baudelet et al., 2015). Significant species differences were observed for some compounds. Useful P2X7 antagonists for human as well as rodent P2X receptors include A438079 (136), A740003 (137), A804598 (138), A839977 (139), AZ1060612 (140), AZ11645373 (141), and GW791343 (142) belonging to different chemical compound classes (see Figure 7). All of the P2X7 antagonists displayed in Figure 7 show potency in the low nanomolar concentration range and high selectivity versus other P2X receptor subtypes.
Several clinical studies have been performed with P2X7R antagonists. Results with AZD9056 (142) and CE-224,535 (145) (Duplantier et al., 2011) in rheumatoid arthritis were published (Keystone et al., 2012; Stock et al., 2012). Both drugs failed to show a clear benefit for the patients. Related, more lipophilic benzamide derivatives (143, 144) were reported to show brain penetration (Wilkinson et al., 2014). These structures indicate that the P2X7 receptor tolerates very bulky residues, like cycloheptyl, adamantyl and related ring systems. Several other P2X7 antagonists that penetrate well into the brain have recently been described, including GSK1482160 (146), which has been prepared in 11C-labeled form to provide a ligand for positron emission tomography (PET) studies (Gao et al., 2015), JNJ-47865567 (147) (Letavic et al., 2013; Bhattacharya et al., 2013), JNJ-42253432 (148) (Lord et al., 2014) and the triazolopyrazinylmethanone 149 (Rudolph et al., 2015). A novel, useful 3H-labeled antagonist radioligand, [3H]JNJ-54232334 (150) has been reported (Lord et al, 2015).
Many novel ligands are now available as pharmacological tool compounds to define action at subtypes of the ARs, P2YRs and P2XRs in the nervous system. We have attempted to focus on the most useful agents in this review. The development of SAR at the adenosine receptors, and to a lesser extent at P2Y and P2XRs, has led to therapeutic concepts and experimental agents for treatment of diseases of the nervous system. Some of these compounds, including A1 and A3 AR agonists, A2A antagonists, P2Y1R and P2Y12R antagonists, and P2X3R, P2X4R and P2X7R antagonists, are potentially useful in disorders of the nervous system, such as chronic pain, neurodegeneration and brain injury. Researchers using these compounds should carefully consider the pharmacological properties as reported in primary literature to avoid their improper application. For example, the selectivity defined for some of the AR agonists and antagonists that were deemed selective for the A1AR or A3AR has been revised with updated pharmacological characterization. Notably, species differences occur frequently leading to a reduction in selectivity ratios in certain species. Also, many of the P2R ligands still lack bioavailability due to charged groups or hydrolytic (either enzymatic or chemical) instability.
The ligand development is continuing; thus, there is reason to expect that subtypes that do not yet have definitive agonist or antagonist ligands will have them in coming years. Although very selective ligands already exist for ARs and there are fewer definitive P2XR and P2YR ligands, the therapeutic potential of selectively modulating these receptors is continuing to gain interest in such fields as cancer, inflammation, pain, diabetes, ischemic protection and many other conditions. X-ray crystallographic structures of the AR and P2YR families have shifted the mode of ligand discovery to structure-based approaches rather than previous empirical approaches. The X-ray structures can be utilized either for in silico screening of chemically diverse libraries for novel ligands or for enhancement of the properties of known ligands by chemical modification. Although X-ray structures of the P2X4R have been reported, there is scant structural information about ligand recognition in this class of trimeric ion channels. There are currently no ligands that are selective for different combinations of P2X subunits. In summary, there are definitive selective agonists and antagonists for all of the ARs and some of the P2YRs, but the pharmacochemistry of P2XRs is still in nascent stages.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.