Animal models for the study of adenosine receptor function


  • R. Yaar,

    1. Departments of Biochemistry and Pharmacology, Whitaker Cardiovascular Institute, Department of Neurology, Boston University School of Medicine, Boston, Massachusetts
    Search for more papers by this author
  • M.R. Jones,

    1. Departments of Biochemistry and Pharmacology, Whitaker Cardiovascular Institute, Department of Neurology, Boston University School of Medicine, Boston, Massachusetts
    Search for more papers by this author
  • J.-F. Chen,

    1. Departments of Biochemistry and Pharmacology, Whitaker Cardiovascular Institute, Department of Neurology, Boston University School of Medicine, Boston, Massachusetts
    Search for more papers by this author
  • Katya Ravid

    Corresponding author
    1. Departments of Biochemistry and Pharmacology, Whitaker Cardiovascular Institute, Department of Neurology, Boston University School of Medicine, Boston, Massachusetts
    • Biochemistry, K225, Boston University School of Medicine, 715 Albany St., Boston, MA.
    Search for more papers by this author


Adenosine receptors represent a family of G-protein coupled receptors that are ubiquitously expressed in a wide variety of tissues. This family contains four receptor subtypes: A1 and A3, which mediate inhibition of adenylyl cyclase; and A2a and A2b, which mediate stimulation of this enzyme. Currently, all receptor subtypes have been genetically deleted in mouse models except for the A2b adenosine receptor, and some have been overexpressed in selective tissues of transgenic mice. Studies involving these transgenic mice indicated that receptor levels are rate limiting, as effects were amplified upon increases in receptor level. The knockout models pointed to clusters of activities related to the physiologies of the cardiovascular and the nervous systems, which are either reduced or enhanced upon specific receptor deletion. Interestingly, the trend of effects on these systems is similar in the A1 and A3 adenosine receptor knockout mice and opposite to the effects observed in the A2a adenosine receptor knockout model. This review summarizes in vitro studies on pathways affected by each adenosine receptor, and primarily focuses on the above in vivo models generated to investigate the physiologic role of adenosine receptors. Furthermore, it illustrates the need for multiple adenosine receptor subtype deficiency studies in mice and the deletion of the A2b subtype. © 2005 Wiley-Liss, Inc.

Classification of adenosine receptors

Initial classification of adenosine receptors into A1 and A2 subtypes was based on the order of potency of ligands, differing responses to ligands, and antagonism by methylxanthines (van Calker et al., 1979; Londos et al., 1980). Initial subclassification of A2-type receptors into A2a and A2b subtypes was also based on pharmacology, specifically, in relation to high affinity A2a adenosine receptor (striatal) and low affinity A2b adenosine receptor (throughout the brain) binding sites in rat brain (Daly et al., 1983). Since then, definitive evidence for the existence of each type of adenosine receptor has been established by molecular cloning and expression studies.

G-protein coupled receptors

Adenosine receptors are all putative seven-pass transmembrane proteins (G-protein coupled) with approximately 21–28 amino acids in each transmembrane region. The N-terminus of the receptor is extracellular and all receptors, with the exception of the A2a adenosine receptor, have a palmitoylation site near the C-terminus (Armstrong et al., 2001). All are glycosylated on the 2nd extracellular loop, although so far, glycosylation appears to have no effect on ligand binding (Piersen et al., 1994). Residues required for coupling to G-proteins may also vary between receptors. Whereas the 3rd intracellular loop and the C-terminus contribute to A1 receptor coupling to G-proteins, only the 3rd intracellular loop has been implicated in coupling of A2a to G-proteins (Tucker et al., 2000). Phosphorylation of intracellular loops also appears to be involved in desensitization of adenosine receptors (Palmer and Stiles, 1997a).

Adenosine as a ligand

Often, more than one type of adenosine receptor is expressed on a single cell type, suggesting possible interactions between different receptors of this family. The concentration of adenosine is likely to be the determining factor in the activation of a receptor type on a particular cell. EC50 values for adenosine at rat A1, A2a, A2b, and A3 receptors, expressed in Chinese hamster ovary cells, have been estimated at 73, 150, 5,100, and 6,500 nM, respectively (Daly and Padgett, 1992; Zhou et al., 1992; Peakman and Hill, 1994). Furthermore, coexisting adenosine receptors with opposing actions on adenylate cyclase activity have been described in a number of cells. For example, the smooth muscle cell line DDT1 MF-2, as well as glomerular and mesangial cells, are reported to express both the A1 and A2a adenosine receptors (Ramkumar et al., 1991; Olivera and Lopez-Novoa, 1992). Cultured porcine coronary artery smooth muscle cells are reported to express the A1, A2a, and A2b adenosine receptors (Mills and Gewirtz, 1990), whereas primary rat vascular smooth muscle cells express A2-type and A3 adenosine receptors (Zhao et al., 1997).

Differences in ligand potency and the need for animal models to study receptor function

Each of the receptors appears to have a similar genomic structure with a single intron inserted into the coding region corresponding to the 2nd intracellular loop. Overall, there is a relatively low level of amino acid homology between different adenosine receptors from a single species, and of the same adenosine receptor between species. This may explain the considerable species differences in ligand potency and selectivity for this class of receptors. For example, although human and sheep receptors are susceptible to antagonism by xanthines and their derivatives, rat, rabbit, and gerbil A3 adenosine receptors are somewhat resistant to these effects. In fact, some xanthine-resistant responses to administered adenosine in tissues from these animals may be due to mast cell (A3 adenosine receptor)-mediated effects (Collis and Brown, 1983; Brackett and Daly, 1991; Martin, 1992; Prentice and Hourani, 1996).

Most adenosine receptor agonists are derivatives of adenosine, which are stabilized by N6 and C2 substitutions of the adenine base, and C5 substitutions of the ribose moiety. 5′-(N-Ethylcarboxamido)adenosine (NECA) is a commonly used agonist that does not discriminate between human receptors, although it demonstrates moderately weaker binding to the rat A3 receptor, as compared to the other rat adenosine receptors (Brown et al., 2001). To be considered selective, agonists and antagonists should ideally differ in potency by at least two orders of magnitude. This is, however, not always the case with adenosine receptor ligands, thus limiting their use for studying specific receptor function in vivo.

Mouse models engineered to overexpress or to harbor a deletion of a specific adenosine receptor have led to significant conclusions on different adenosine receptor function. This review summarizes the distribution and signaling properties of each of the adenosine receptors, and focuses on the role of each receptor as revealed from studies with knockout and transgenic mice.


A1AR distribution and signaling

The A1 adenosine receptor has been localized to human chromosome 1q32.1 (Megson et al., 1995) and this region maps to mouse chromosome 1 (LocusLink; Although the primary sequence for this receptor is more highly conserved between species than that of other adenosine receptors, there are still significant differences in ligand binding and desensitization of A1 receptors (Ramkumar et al., 1991; Zhou et al., 1992; Linden et al., 1993; Nie et al., 1997; Palmer and Stiles, 1997a). Whereas the human and rat A1 receptors demonstrate greater affinity for the agonist NECA than S-N6-(phenylisopropyl)-adenosine (S-PIA), the opposite is true of bovine receptors (Zhou et al., 1992). Also, despite its high homology to A1 receptors in other species, the guinea pig A1 receptors displays a unique pharmacological profile: high affinity for the A1-selective antagonist 1,3-dipropyl-8-cyclopentylxanthine (DPCPX), yet very low affinity for some A1-selective agonists such as N6-cyclopentyladenosine (CPA), N6-cyclohexyladenosine (CHA), and R-N6-(phenylisopropyl)-adenosine (R-PIA) (Meng et al., 1994).

The 326 amino acid A1 adenosine receptor is widely distributed and particularly prevalent in the central nervous system, with high levels in the cerebral cortex, hippocampus, cerebellum, thalamus, brain stem, and spinal cord of rat (Reppert et al., 1991; Dixon et al., 1996). Furthermore numerous peripheral tissues express the A1 adenosine receptor, including vas deferens, testis, white adipose tissue, stomach, spleen, pituitary, adrenal, heart, aorta, liver, eye, and bladder (Reppert et al., 1991; Dixon et al., 1996). Low levels are found in lung, kidney, and small intestine (Reppert et al., 1991; Stehle et al., 1992; Dixon et al., 1996). A similar distribution has been described for human A1 receptors (Ren and Stiles, 1994).

Activation of A1 receptors results in decreased adenylate cyclase activity through activation of pertussis toxin-sensitive Gi-proteins (van Calker et al., 1978; Londos et al., 1980), although presynaptic A1 receptors are reported to be at least partially resistant to this toxin (Fredholm et al., 1989; Hasuo et al., 1992). A1 adenosine receptors expressed in Chinese hamster ovary K1 cells increase levels of calcium and inositol phosphate upon activation, although it is unclear whether this demonstrates a direct effect or a consequence of cyclic 3′-5′-adenosine monophosphate (cAMP) elevation (Iredale et al., 1994; Megson et al., 1995).

In cardiac muscle and neurons, A1 receptors can also activate pertussis toxin-sensitive K+ channels, as well as KATP channels. Coupling to K+ channels in supraventricular tissue is responsible for the bradycardic effect of adenosine on heart function (Belardinelli et al., 1995; Bunemann and Pott, 1995; Ito et al., 1995). KATP channel coupling has been demonstrated in rat and guinea-pig ventricular myocytes (Kirsch et al., 1990; Ito et al., 1994), porcine coronary arteries (Merkel et al., 1992; Dart and Standen, 1993), rabbit heart (Nakhostine and Lamontagne, 1993), and rat cerebral cells (Cass et al., 1991; Klapstein and Colmers, 1992), and is purported to mediate reductions in action potential duration as well as vasodilation.

Biological significance: studies in genetically engineered animals

A1 adenosine receptors play a role in a diverse array of biological effects. These receptors are found both pre- and post-synaptically, and are implicated in the xanthine-sensitive sedative, anxiolytic, and locomoter depressant effects of adenosine in the central nervous system (Nikodijevic et al., 1991; Higgins et al., 1994; Jain et al., 1995; Malhotra and Gupta, 1997). Other studies have demonstrated that the chronotropic, dromotropic, and inotropic depression of the heart in response to systemic adenosine administration is attributed to A1 receptors coupled to K+ channels (Olsson and Pearson, 1990). Although there is some evidence for a direct effect of A1 adenosine receptor activation on vascular tone (Stoggall and Shaw, 1990; Szentmiklosi et al., 1995), most vascular effects result from the modulation of neurotransmitter release (Rubino et al., 1993; Goncalves and Queiroz, 1996).

Abundant evidence reveals an important role for A1 adenosine receptors in protection of the heart (Sakamoto et al., 1995; Mizumura et al., 1996; Stambaugh et al., 1997), lung (Neely and Keith, 1995), and brain (Neely and Keith, 1995) from ischemia-reperfusion injury. Furthermore, these studies recently have been supported by the demonstration that cardiac overexpression of A1 receptors in transgenic mice results in substantial protection from ischemia-reperfusion injury (Matherne et al., 1997; Gauthier et al., 1998; Yang et al., 2002).

Mice expressing approximately 100-fold higher levels of A1 receptors in the heart, compared to wild type mice, have been generated by coupling cDNA for the rat A1 receptor with the cardiac-specific alpha myosin heavy chain promoter (Matherne et al., 1997). Perfused hearts that were isolated from transgenic animals have a lower basal heart rate compared to control mice, suggesting that endogenous adenosine levels are sufficient to activate the overexpressed receptor. When isolated hearts were paced, myocardial function tests, including response to catecholamines, were comparable between transgenic and control mice. During ischemia, however, transgenic mice showed a significantly increased time of contracture as well as improved functional recovery during reperfusion (Matherne et al., 1997; Gauthier et al., 1998). The A1 adenosine receptor cardiac overexpression mice also show an improved bioenergetic state [ATP/(ADP.Pi)] and better recovery of phosphocreatine during reperfusion (Gauthier et al., 1998). Furthermore, although preconditioning improved functional recovery in hearts of control mice, it did not change functional recovery in transgenic hearts (Morrison et al., 2000). These findings suggest that the A1 adenosine receptor mediates many of the effects associated with cardiac preconditioning and that the benefits of the A1 adenosine receptor in the context of ischemia-reperfusion injury are limited by endogenous levels of the receptor.

Generation of mice deficient in A1 receptors has also supported previous research as to the role of the A1 adenosine receptor in central nervous system and kidney physiology (Brown et al., 2001). Mice lacking A1 receptors do not demonstrate tubuloglomerular feedback in response to increased flow in the loop of Henle when compared with normal mice, although other measures of kidney function are normal (Brown et al., 2001). Similar to the phenotype observed in A1-deficient animals, A1 adenosine receptor antagonists have been used in experimental animals and during clinical trials that have resulted in diuresis and natriuresis in vivo (Wilcox et al., 1999; Gottlieb et al., 2002). These studies demonstrate that the A1 adenosine receptor is necessary for the tubuloglomerular response to increased flow rate and subsequently impacts renal water and sodium retention. Furthermore, plasma renin activity is increased in A1-deficient mice compared to their wild type littermates (Brown et al., 2001), suggesting that adenosine regulates renin release via A1 receptors.

Mice lacking functional A1 adenosine receptors show signs of increased anxiety and hyperalgesia (Johansson et al., 2001). Adenosine-mediated inhibition and theophylline-mediated augmentation of glutamatergic neurotransmission from hippocampal slices is lost in A1 adenosine receptor deficient mice. The response to these agents is halved in mice hemizygous for the A1 adenosine receptor gene, suggesting a lack of a receptor reserve in these cells. Analgesic effects of intrathecal adenosine are lost in A1 receptor deficient mice when compared to normal mice, and thermal hyperalgesia is observed (Johansson et al., 2001). Furthermore, hypoxia-associated decrease in neuronal activity is less pronounced in A1 adenosine receptor deficient null mice, as is recovery of neuronal activity after hypoxia (Brown et al., 2001). Thus, although the A1 adenosine receptor may not play an essential role in the normal physiology of nervous tissue, this receptor has important effects in pathophysiologic conditions such as noxious stimulation and hypoxia. Major consequences of deleting the A1 adenosine receptor in a mouse model are summarized in Figure 1.

Figure 1.

Consequences of A1 adenosine receptor (A1AR) deficiency in mice.


A2aAR distribution and signaling

The human A2a adenosine receptor gene has been localized to chromosome 22q11.23 (MacCollin et al., 1994; Dubey et al., 1996; Deckert et al., 1997) and this region is localized to mouse chromosome 10 (LocusLink; At 410 amino acids (409 in mouse and rat), this receptor is larger than the other adenosine receptors, a characteristic that is wholly attributable to an extended C-terminus of no known function (Piersen et al., 1994; Jacobson et al., 1997). Expression of the A2a adenosine receptor is high in spleen, thymus, leukocytes, blood platelets, and in the central nervous system (Meng et al., 1994; Peterfreund et al., 1996). Intermediate levels of expression are found in the heart, lung, and blood vessels (Meng et al., 1994; Peterfreund et al., 1996). Within the central nervous system, the A2a adenosine receptor is highly expressed in the striatum, nucleus accumbens, and olfactory tubercle (Santicioli et al., 1993).

The A2a adenosine receptor is coupled to cholera toxin-sensitive Gs-proteins, and activation of A2a receptors results in increased adenylate cyclase activity (Palmer and Stiles, 1995). Although Gs appears to be the major G-protein associated with A2a receptors in the periphery, there is a relative deficiency of Gs in the striatum, where A2a receptor concentrations are highest. There is now evidence that striatal A2a receptors mediate their effects predominantly through activation of Golf (Fredholm et al., 2000).

Biological significance: studies in genetically engineered animals

Like the A1 adenosine receptor, the A2a adenosine receptor has been implicated in a wide range of biological events. The A2a adenosine receptor is expressed in many vascular beds and is associated with vasodilation in rat aorta and bovine coronary artery (Conti et al., 1993). Vasodilation in some other vascular beds, however, appears to be mediated by A2b receptors (Rubino et al., 1993; Szentmiklosi et al., 1995). The A2a adenosine receptor is also expressed in platelets, and has been shown to inhibit aggregation by increasing intracellular cAMP levels (Varani et al., 2000). Chronic consumption of caffeine, a nonselective adenosine receptor antagonist, increases platelet A2a adenosine receptor binding sites in humans, and sensitizes platelets to 2-hex-1-ynyl-5′-N-ethylcarboxamidoadenosine (HENECA), a nonselective adenosine receptor agonist derived from NECA (Varani et al., 2000).

The A2a subtype was the first adenosine receptor to be genetically deleted in a murine model (Ledent et al., 1997). Two A2a adenosine receptor knockout mouse lines in three different genetic backgrounds [CD1, (Ledent et al., 1997); congenic C57BL/6 (Chen et al., 2001a) and pure 129-Steel (Chen et al., 1999)] were developed. These animals demonstrate several central nervous system disturbances, including decreased exploratory activity, increased aggressiveness and hypoalgesia (Ledent et al., 1997), compensatory alteration in spinal cord opioid receptors (Bailey et al., 2002), attenuated psychostimulant responses (Chen et al., 2000, 2003), reduced alcohol sensitivity (Naassila et al., 2002), and reduced alcohol withdrawal-induced seizure (El Yacoubi et al., 2001). Exploratory activity in mice is generally increased by caffeine administration, however, the opposite effect was observed in A2a-null mice which suggested that caffeine-dependent psychostimulation is mediated by the A2a receptor (El Yacoubi et al., 2001). Furthermore, A2a-deficient mice on a CD1 genetic background have increased blood pressure, heart rate, and platelet aggregation (Ledent et al., 1997). A2a-deficient mice on either a mixed 129-Steel × C57BL/6 (Chen et al., 1999) or congenic C57BL/6 (Day et al., 2003) genetic background, however, did not display blood pressure differences.

A potential mechanism for A2a-dependent signaling in the brain is via the dopaminergic system. It was shown that A2a adenosine receptors are co-expressed with D2 dopamine receptors (D2Rs) in the striatum (Schiffmann et al., 1991; Fink et al., 1992; Hettinger et al., 2001) and that heterodimerization of A2a and D2 receptor subtypes inhibit D2 receptor function (Fuxe et al., 1998). In SH-SY5Y human neuroblastoma cells, A2a-D2 receptor heteromeric complexes undergo co-aggregation and co-internalization upon long-term exposure to A2a adenosine receptor or D2R agonists (Hillion et al., 2002). These results strongly support the notion that intra-membrane interaction between A2a adenosine receptors and D2Rs is antagonistic. Alternatively, a different receptor interaction model proposed opposing but independent action of A2a and D2R by complementary pharmacological and genetic A2a knockout studies (Svenningsson et al., 1999). For example, A2a adenosine receptor agonists and antagonists produce behavioral and cellular functions in D2R-null mice that are similar to their wild type counterparts thus suggesting a D2-independent mechanism (Aoyama et al., 2000; Zahniser et al., 2000; Chen et al., 2001b). Moreover, A2a activation (apparently independent of D2R) is required for dopaminergic function since dopamine-mediated cellular response such as DARPP-32 phosphorylation and immediately early gene expression in the striatum was abolished in the A2a-deficient mice (Svenningsson et al., 2000; Dassesse et al., 2001). These results suggest that A2a adenosine receptors are capable of D2R-dependent and independent signaling.

Due to the molecular interaction between dopamine and A2a adenosine receptors and because A2a adenosine receptors are primarily expressed in the striatum, A2a antagonists were explored as potential therapeutics in the treatment of Parkinson's disease (PD). It was demonstrated that A2a receptor blockade or genetic depletion of the A2a receptor attenuates dopaminergic neurotoxicity in the MPTP-model of PD (Chen et al., 2001a). Moreover, the non-specific A2a antagonist caffeine also reduced MPTP-induced dopaminergic neurodegeneration (Chen et al., 2001a). This result provides a neurobiological basis for the recent strong epidemiological evidence showing the inverse relationship between caffeine consumption and risk of developing PD in two large prospective studies (Ross et al., 2000; Ascherio et al., 2001). Furthermore, A2a antagonists such as KW6002 have been proposed (Dixon et al., 1997; Schwarzschild et al., 2002), and have shown potential in a recently completed clinical phase II trial as a novel treatment for PD (Bara-Jimenez et al., 2003; Hauser et al., 2003). These findings support the prospect of using A2a antagonists as a novel therapeutic strategy that not only provides symptomatic relief but also decelerates the neurodegeneration of dopaminergic cells in PD patients (Schwarzschild et al., 2002).

The role of the A2a adenosine receptor in neuronal cell death is not limited to dopaminergic systems. Pharmacological and genetic A2a adenosine receptor deficiency studies have demonstrated that A2a inactivation protects against neuronal cell death induced by ischemia (Phillis, 1995; Monopoli et al., 1998; Chen et al., 1999), excitototoxicity (such as quinolinic acid-induced) (Jones et al., 1998; Popoli et al., 2002), the mitochondrial toxins 3-NP (an animal model of Huntington's disease) (Blum et al., 2003; Day et al., 2003), and β-amyloid aggregation (key neuropathological changes in Alzheimer's disease) in cultured cerebral granular cells (Dall'lgna et al., 2003). A broad neuroprotective effect supports the notion that endogenous adenosine acting at A2a adenosine receptors may in fact exacerbate brain tissue damage. It is interesting to speculate that adenosine-mediated neuroprotection is dependent on a delicate balance of opposing A1 and A2a adenosine receptor activation.

Other studies have shown that A2a adenosine receptor agonists result in neuroprotection in some experimental conditions, including cerebral hemorrhage injury (Mayne et al., 2001) and ischemia-reperfusion injury in spinal cord (Cassada et al., 2002). This A2a agonist-mediated cellular protection is particularly evident in peripheral tissues (Linden, 2001). For example, a newly developed A2a agonist, ATL-146e has been shown to protect against ischemic renal injury (Okusa et al., 1999, 2000). This A2a adenosine receptor agonist-induced tissue protection has been attributed to A2a receptor modulation of inflammation during tissue injury. A recent study using A2a-null mice (Armstrong et al., 2001; Ohta and Sitkovsky, 2001) has provided evidence supporting the notion that extracellular adenosine acting at A2a receptors inhibits inflammation (Cronstein et al., 1990). Subthreshold doses of inflammatory stimuli induce extensive tissue damage, as well as more prolonged, and higher levels of pro-inflammatory cytokines in mice lacking the A2a adenosine receptor than in wild type mice. This suggests a critical role for this receptor in attenuation of inflammation (Armstrong et al., 2001; Ohta and Sitkovsky, 2001). Thus extracellular adenosine acting at A2a adenosine receptors results in complex actions in various tissues, and tissue protection by adenosine can only be achieved by targeting the A2a receptors in specific tissues and cellular elements. Major consequences of deleting the A2a adenosine receptor in a mouse model are summarized in Figure 2.

Figure 2.

Consequences of A2a adenosine receptor (A2aAR) deficiency in mice.


A2bAR distribution and signaling

The A2b adenosine receptor has been localized to human chromosome 17p12-11.2 (Jacobson et al., 1995), and this region maps to mouse chromosome 11 (LocusLink; The A2b adenosine receptor is coupled to adenylate cyclase-stimulatory G-proteins, but has also been reported to activate phospholipase C in human mast cells (Feoktistov and Biaggioni, 1995). The receptor is expressed widely, but generally at very low levels. Reverse transcription coupled to the polymerase chain reaction (RT-PCR) and in situ hybridization of rat tissue demonstrates low levels of A2b receptor in all brain regions tested (Stehle et al., 1992; Dixon et al., 1996), whereas, by Northern blot, high levels can be found in the cecum, large intestine, and urinary bladder. Lower levels of the A2b adenosine receptor can be detected in spinal cord, lung, vas deferens, and pituitary (Stehle et al., 1992). Receptor message has recently been described in a variety of skin cells (Le Poole et al., 1999). There are confirmatory reports of a similar profile of expression in human and mouse using an A2b adenosine receptor antibody (Puffinbarger et al., 1995).

Like the A2a adenosine receptor, the A2b adenosine receptor is coupled to Gs and mediates some of its effects by activating adenylate cyclase (Brackett and Daly, 1994; Peakman and Hill, 1994). There is a considerable amount of evidence suggesting that activation of phospholipase C, through Gq proteins, mediates many of the important functions of A2b receptors. Specifically, NECA, a nonselective adenosine receptor agonist, increased inositol phosphate formation in human mast cell line MHC-1, whereas CGS 21680, an A2a receptor-selective agonist, could not (Feoktistov and Biaggioni, 1995). This effect was not sensitive to cholera or pertussis toxin, and was antagonized by enprofylline, an A2b receptor-selective antagonist (Feoktistov and Biaggioni, 1995).

Biological significance: studies in genetically engineered animals

Lack of specific agonists for the A2b adenosine receptor is a major reason why so little is known about the functional significance of this receptor. Despite this limitation, the A2b adenosine receptor has been implicated in several biological events. This receptor plays a role in mediating vasodilation in some vessels, such as guinea pig aorta (Martin, 1992), rat renal artery (Abbracchio et al., 2001), and dog coronary artery (Balwierczak et al., 1991), while vasodilation in other vessels is mediated by A2a receptors (Conti et al., 1993). Inhibition of growth of rat aortic smooth muscle cells has been achieved through selective A2b receptor activation, and appears to be secondary to blockade of the MAPK pathway (Dubey et al., 1996, 2000). Furthermore, this receptor may play a role in mediating allergic or inflammatory disorders. It has been found in mouse bone marrow-derived mast cells (Marquardt et al., 1994), in human mast cells (Feoktistov and Biaggioni, 1996), and in canine BR mastocytoma cells, where it mediates degranulation (Auchampach et al., 1997a). Enprofylline is known to have therapeutic anti-asthmatic effects (Feoktistov and Biaggioni, 1996). Generation of A2b adenosine receptor knockout mice and transgenic mice over expressing this receptor in selected tissues will further enhance our understanding of the various functions of this receptor.


A3AR distribution and signaling

The A3 adenosine receptor has been localized to human chromosome 1p21-p13 (Atkinson et al., 1997), and this region maps to mouse chromosome 3 (LocusLink; The A3 adenosine receptor is the only adenosine receptor cloned prior to its pharmacologic identification. Originally isolated as an orphan receptor from rat testis (Meyerhof et al., 1991), the A3 adenosine receptor was later cloned from rat striatum and expressed and characterized in COS-7 and Chinese hamster ovary cells (Zhou et al., 1992).

A3 receptor activation inhibits adenylate cyclase activity (Zhou et al., 1992) by activation of pertussis toxin-sensitive Gia2,3-proteins (Palmer and Stiles, 1995). In the rat mast cell line RBL-2H3 (Ali et al., 1990; Ramkumar et al., 1993) and rat brain (Abbracchio et al., 1995), A3 receptors can stimulate phospholipase C by activation of Gq/11. In contrast to other adenosine receptors, recombinant rat and human A3 receptors desensitize within minutes of agonist exposure by uncoupling from G-proteins. Receptor phosphorylation by G-protein receptor kinases appears to be responsible for this event (Palmer and Stiles, 1995, 1997b), and similar results have been described in RBL-2H3 cells (Ali et al., 1990; Ramkumar et al., 1993). The C-terminus has been identified as the most likely site of phosphorylation for desensitization (Palmer et al., 1996).

While the A3 adenosine receptor is widely distributed in most animals, there exist pronounced differences in expression levels between species. The A3 receptor has been detected in testis, lung, kidney, placenta, heart, brain, spleen, liver, uterus, bladder, jejunum, aorta, proximal colon, and eye of rat, sheep, and human, although message has not been detected in skin or skeletal muscle (Zhou et al., 1992; Salvatore et al., 1993; Linden, 1994; Rivkees, 1994; Dixon et al., 1996). Rat testis has particularly high concentrations of A3 receptor mRNA, localized in spermatocytes and spermatids (Linden et al., 1993; Salvatore et al., 1993), with moderate levels in lung (Dixon et al., 1996). The highest levels of human A3 adenosine receptor message have been found in lung and liver, with lower levels in aorta and brain (Salvatore et al., 1993).

Although high levels of the A3 adenosine receptor are found in germ cells of the rat testis, no specific function has been ascribed to this expression and mice lacking A3 receptors reproduce normally under laboratory conditions (Salvatore et al., 2000). Several groups have detected low levels of A3 receptor message in the brain in all species examined, and there is some evidence of higher message levels in the hypothalamus and thalamus (Linden et al., 1993; Dixon et al., 1996; Zhao et al., 2002). Despite low levels of central nervous system expression, intraperitoneal injection of N6-(3-iodobenzyl)adenosine-5′-N-methyluronamide (IB-MECA), a selective A3 receptor agonist, depresses locomoter activity in mice (Jacobson et al., 1993), suggesting role for this receptor in depression of motor activity.

Biological significance: studies in genetically engineered animals

The A3 adenosine receptor has been implicated in mediating allergic responses. This receptor facilitates the release of allergic mediators, such as histamine, in mast cells (Ramkumar et al., 1993), and systemic infusion of IB-MECA has been shown to cause scratching in mice that is prevented by coadministration of histamine antagonists (Jacobson et al., 1993). Moreover, A3 adenosine receptor message has been detected, by in situ hybridization of human lung tissue, in mesenchymal cells and eosinophils of the lung (Walker et al., 1997). Interestingly, mean A3 adenosine receptor transcript is elevated in lung tissue from individuals with airway inflammation and in peripheral eosinophils of atopic individuals (Walker et al., 1997). This suggests either an upregulation of A3 receptors in the presence of inflammation or exacerbation of inflammation by elevated A3 receptor levels.

There is some discrepancy in the literature regarding the role of the A3 adenosine receptor in mediating apoptotic events. In human eosinophils (Kohno et al., 1996a) and human promyelocytic HL-60 cells (Kohno et al., 1996b; Yao et al., 1997) A3 adenosine receptors seem to be involved in apoptosis at relatively high concentrations (>10 μM) of agonist, or by chronic activation of the receptor. Interestingly, A3 receptor antagonists such as MRS1191 (0.5 μM), also induce apoptotic cell death in human HL-60 leukemia and U-937 lymphoma cell lines, an effect that is opposed by low concentrations (10 nM or 1 μM) of Cl-IB-MECA (Kohno et al., 1996b; Yao et al., 1997). Furthermore, in RBL-2H3 cells, IB-MECA (1 μM) potently blocks apoptosis induced by ultraviolet irradiation (Gao et al., 2001). It is possible that modulation of apoptosis by A3 receptor activation is dependent both on the cell type involved and/or the type of activation.

A role for the A3 adenosine receptor in mediating control of the cell cycle has been reported (Neary et al., 1998). The A3 adenosine receptor can activate ERK1/2 in human fetal astrocytes. Moreover, A3 receptor activation has been implicated in inhibition of tumor growth both in vitro and in vivo (Fishman et al., 2000a,b, 2001; Bar-Yehuda et al., 2001; Ohana et al., 2001). Interestingly, attempts to overexpress the A3AR in a transgenic mouse model resulted in embryonic lethality (Zhao et al., 2002). These studies open the possibility for using selective activation of A3 receptors as a tool in the treatment of cancer.

Several reports suggest a role for the A3 adenosine receptor in preconditioning the heart for ischemic injury (Armstrong and Ganote, 1994, 1995; Liu et al., 1994; Stambaugh et al., 1997; Auchampach et al., 1997b). Preconditioning of rabbit cardiomyocytes is blocked by the nonselective A1/A3 receptor agonist N6-2-(4-aminophenyl)ethyladenosine (APNEA), but not by the A1 receptor selective agonist R-PIA (Armstrong and Ganote, 1995). Other studies, however, have shown that maximal preconditioning requires both A1 and A3 adenosine receptor activation (Wang et al., 1997). Activation of A3 adenosine receptors on several cell types, including RBL-2H3 cells, endothelial cells, cardiomyocytes, and smooth muscle cells upregulates cellular antioxidant systems that may play a role in the cytoprotective actions of A3 receptor activation during ischemia (Maggirwar et al., 1994).

Mice carrying a genetic deletion of the A3 adenosine receptor have been described in several reports. The ability of Cl-IB-MECA to potentiate antigen-dependent degranulation of mast cells, as measured by hexosaminidase release, was lost in mice lacking A3 receptors, when compared to normal mice (Salvatore et al., 2000). Furthermore, attenuation of lipopolysaccharide-induced tumor necrosis factor alpha production was decreased in mice lacking A3 receptors, as compared to control mice (Salvatore et al., 2000). Finally, cutaneous vasopermeability is associated with activation and subsequent degranulation of mast cells, and induction of cutaneous vasopermeability by adenosine or inosine, as measured by extravasation of plasma protein, is completely lost in mice lacking a functional A3 adenosine receptor (Tilley et al., 2000).

One of the most well known actions of adenosine is the ability to dilate most vascular beds, potentially by interacting with adenosine receptors and modulating cAMP levels in the vasculature. Although there is no significant change in levels of A1 or A2 receptors in aortas of A3 receptor deficient mice, steady-state levels of cAMP are elevated when compared to control mice. Studies of blood pressure response to intravenous adenosine injection show a significantly larger drop in blood pressure in mice lacking the A3 adenosine receptor when compared to control mice (Zhao et al., 2000). Despite these effects, which are induced upon the deletion of the A3 adenosine receptors, the development of atherosclerosis or response to injury of the femoral artery are similar to these in wild type mice (Jones et al., 2004). This was followed after cross breeding the A3 adenosine receptor null mice with ApoE null mice and following protocols for atherosclerosis and vascular injury as in Cayatte et al. (2001).

Two groups have analyzed ischemia-reperfusion injury in mice lacking a functional A3 adenosine receptor, using an intact heart model. Interestingly, ischemia-reperfusion injury of mice lacking A3 adenosine receptors results in a significant reduction in infarct size, compared to wild type mice. No difference in infarct size is seen after preconditioning (Guo et al., 2001). Hearts from mice lacking the A3 adenosine receptor also show improved recovery of left ventricular developed pressure and increased tissue viability when compared to control mice, in an isovolumic Langendorff perfusion model (Cerniway et al., 2001). Finally, neutrophil infiltration of damaged myocardium is reduced in mice lacking A3 receptors, suggesting a role for these receptors in exacerbating ischemia-reperfusion injury through neutrophil chemotaxis (Guo et al., 2001). Major consequences of deleting the A3 adenosine receptor in a mouse model are summarized in Figure 3.

Figure 3.

Consequences of A3 adenosine receptor (A3AR) deficiency in mice.


The availability of ligands and antagonists selective for different adenosine receptors allowed verification of different receptor functions mainly via in vitro studies. The usage of these agonists in vivo has often raised concerns about their selectivity when employed at dosages needed to elicit effects. Hence, the generation of transgenic mice overexpressing the receptors or knockout mice deleted of a selected adenosine receptor has proven valuable. The former animal models indicated that receptor levels are rate limiting, as effects were amplified upon increases in receptor level. The knockout models pointed to clusters of related activities, which are either reduced or enhanced upon specific receptor deletion. Interestingly, thus far two major systems have been identified to be affected by adenosine receptor deletion: the cardiovascular as well as the nervous system. As illustrated in Figures 1 and 3, the trend of effects on these systems is similar in the A1 and A3 adenosine receptor knockout mice and opposite to the observed in the A2a adenosine receptor knockout model. Future generation of A2b adenosine receptor null mice will complete this line of analysis. Also, cross-breeding different adenosine receptor knockout mice to create combinations of deletions should prove useful in elucidating the importance of this receptor family in survival, reproduction, and normal physiology.