Mechanisms of induction of adenosine receptor genes and its functional significance

Abstract

Adenosine is a metabolite generated and released from cells, particularly under injury or stress. It elicits protective or damaging responses via signaling through the adenosine receptors, including the adenylyl cyclase inhibitory A₁, and A₃, and the adenylyl cyclase stimulatory A_2A and A_2B. Multiple adenosine receptor types, including stimulatory and inhibitory, can be found in the same cell, suggesting that a careful balance of adenosine receptor expression in a particular cell is necessary for a specific adenosine-induced response. This balance could be controlled by differential expression of the adenosine receptor genes under different stimuli. Here, we have reviewed an array of studies that have characterized basal or induced expression of the adenosine receptors and common as well as distinct mechanisms of effect, in hopes that ongoing studies on this topic will further elucidate detailed mechanisms of adenosine receptor regulation, leading to potential therapeutic applications.

Adenosine: generation and its targets

Adenosine is considered a regulatory metabolite and signals via binding to the four adenosine receptors (AR) classified by their ability to inhibit or stimulate adenylyl cyclase – the Gα_i coupled A₁AR and A₃AR, and the Gα_s coupled A_2AAR and A_2BAR (Fredholm et al., 2001; Hasko and Cronstein, 2004). Adenosine is generated in response to hypoxia, injury, or inflammation and it's signaling occurs in processes such as: increasing the oxygen supply/demand ratio, ischemic pre- and postconditioning, anti-inflammatory responses, and angiogenesis (Linden, 2005). Adenosine is the product of adenosine 5'-triphosphate (ATP) catabolism. Intracellular levels of ATP are in the millimolar range, therefore compromised membrane integrity or conditions of inflammation, ischemia or hypoxia will increase the extracellular concentration of ATP significantly (Bours et al., 2006). Additionally, processes such as nerve stimulation (Rump et al., 1996), hypotonic stress (Van der Wijk et al., 1999) and mechanical stress (Sauer et al., 2000) induce cells to release ATP. Once outside the cell, a series of membrane-bound enzymes rapidly degrade ATP. Ecto-nucleoside triphosphate dephosphorylases (E-NTPDases) such as CD39, catalyze the dephosphorylation of ATP to adenosine 5'-diphosphate (ADP) and ADP to adenosine 5'-monophosphate (AMP) (Kaczmarek et al., 1996; Wang and Guidotti, 1996). The family of enzymes referred to as ecto-pyrophosphatase/phosphodiesterases (NPPs) hydrolyse ATP directly to AMP (Goding et al., 2003). The last step of converting extracellular AMP to adenosine is performed by ecto-5'-nucleotidase referred to as CD73 (Eckle et al., 2007). It is interesting to note that CD39, and adenosine receptors A₁ and A_2B associate with caveolae, suggesting that these adenosine-generating enzymes may be localized to the same regions on the cell surface where adenosine receptors are present, allowing for newly-formed adenosine to quickly bind its receptors (Kittel et al., 1999; Lasley et al., 2000; Gines et al., 2001; Sitaraman et al., 2002; Matsuoka and Ohkubo, 2004; Kittel et al., 2005). Similarly, conditions such as hypoxia inhibit the enzyme adenosine kinase, thus potentiating the increase in extracellular adenosine (Decking et al., 1997). Intracellular adenosine can be transported outside the cells by bi-directional adenosine transporters of two families: equilibrative transporters and concentrative transporters. Equilibrative transporters, referred to as ENTs, can function via facilitated diffusion (such as ENT1 an ENT2) or cation-dependent transport (such as ENT3 and ENT4) to move adenosine into the extracellular space (Baldwin et al., 2004). Concentrative transporters (CNTs) are less abundant than ENTs, and function in a sodium-dependent manner to transport adenosine, against the concentration gradient, into the cell (Gray et al., 2004).

The affinity of adenosine for its different receptors and the downstream effects of ligand-receptor interaction vary. Basal body fluid concentrations of adenosine are able to stimulate the A₁AR, A₃AR and A_2AAR. However, pathological conditions such as ischemia or tissue damage cause the concentration of adenosine to increase enough to allow for stimulation of the low-affinity A_2BAR (Fredholm, 2007). Most research has focused on the downstream functions of adenosine receptor activation, while the control of adenosine receptor expression at the gene or protein level has received less attention.

Overview of functional roles of adenosine receptors in different tissues and cells

The A₁AR

The rat A₁AR gene was cloned in 1992 (Mahan et al., 1991) and subsequent studies have shown it to have a protective effect on cardiac tissue during stress/injury. (Mahan et al., 1991). In conjunction with pharmacological activation, when the A₁AR was over-expressed in cardiac tissues of mice subjected to ischemia/reperfusion (I/R) the mice showed a protective response. The rate-limiting component of this protection was the amount of A₁AR expressed (Matherne et al., 1997). One result of I/R or myocardial ischemia is apoptosis of cardiac cells, and programmed cell-death contributes to myocardial malfunction (Gottlieb and Engler, 1999; Sam et al., 2000). Mice overexpressing A₁AR in cardiac tissue have a decrease in caspase 3 activation and therefore less apoptosis after I/R, suggesting the protective effects of A₁AR in I/R function via reducing apoptosis of cardiac cells (Regan et al., 2003). Studies in A₁AR knockout mice further solidified this point as hearts from A₁AR knockout mice had reduced recovery and systolic blood pressure after an ischemic event (Reichelt et al., 2005).

Additionally, A₁AR has been found to ameliorate the inflammatory response, which may also contribute to its protective qualities. A₁AR knockout mice were shown to be susceptible to the hyper-acute inflammatory response that results from sepsis (Gallos et al., 2005). Preconditioning via activation of the A₁AR prior to an ischemic event in the kidney yields protective effects by reducing the inflammatory response as well as cell death resulting from the injury (Lee and Emala, 2000; Lee et al., 2004).

Though A₁AR’s role in cardiovascular health has been the most studied, other roles have more recently been found. An A₁AR agonist, WRC-0571, was recently shown to promote angiogenesis. It also stimulated vascular endothelial growth factor (VEGF) release from mononuclear cells, illustrating the complex interplay between different cell types and A₁AR activation (Clark et al., 2007). Additionally, the A₁AR is highly expressed in the brain and is involved in synaptic transmission in the hippocampus (Cunha-Reis et al., 2007). Anticonvulsant effects of adenosine were shown to function through A₁AR activation (Kochanek et al., 2006). An A₁AR-specific agonist, FR194921, improved memory defects and reduced anxiety in rats, suggesting a role for A₁AR in anxiety disorders in humans (Maemoto et al., 2004). Furthermore, in muscle, an A₁AR-specific agonist, 2-chloro-N(6)-cyclopentyladenosine (CCPA), caused an increase in glucose transport induced by insulin signaling, but did not change the overall amount of glucose uptake (Thong et al., 2007). The protective role of A₁AR involves signaling through its associated G_i protein and downstream effectors including the Akt/mitogen-activated protein kinase (MAPK) and extracellular signal-regulated kinase (ERK)1/2 pathways, although the pathway appears to be cell-type specific. In the mouse heart, the p38 mitogen-activated protein kinase (MAPK) signaling pathway is responsible for protection from ischemia (Zhao et al., 2001). Endogenous adenosine (Haq et al., 1998) and A₁AR activation via the agonist CCPA (Dana et al., 2000) activates p38 MAPK in ischemic rat and rabbit hearts, respectively. Additionally, a p38 MAPK inhibitor abolished the cardio-protective effect of adenosine in pig hearts (Yoshimura et al., 2004). In renal tubules, A₁AR activation protects from H₂O₂ induced-apoptosis via phospho-protein kinase C (PKC) and ERK1/2 signaling, with ultimate activation of heat shock protein 27 (Hsp27) (Lee et al., 2007). In astrocytes, the A₁AR protects from hypoxia/ischemia-induced apoptosis by signaling through phosinositide 3-kinase (PI3K) and ERK1/2 MAPK pathways (Ciccarelli et al., 2007).

The A₃AR

The A₃AR gene was first cloned in 1992 from rat brain cDNA libraries (Zhou et al., 1992). With the use of a knockout model, the A₃AR was found to play a role in the regulation of steady-state levels of cAMP in cells expressing A₃AR under basal conditions. Additionally, the A₃AR knockout mouse demonstrated changes in blood pressure due to adenosine, in part, via A₃AR signaling (Zhao et al., 2000). The A₃AR has similar cardioprotective effects as the A₁AR under ischemic or hypoxic conditions (Liang and Jacobson, 1998; Hochhauser et al., 2007). Agonists of both A₁AR and A₃AR protected cardiomyocytes from deleterious effects of hypoxia (Safran et al., 2001). In dogs, an A₃AR agonist IB-MECA reduced myocardial damage when administered during reperfusion, suggesting its potential in a clinical setting (Auchampach et al., 2003). In lung pathologies of ischemia, such as during a lung transplant, stimulation of A₃AR was shown to protect the lung and reduce ischemia-related apoptosis (Rivo et al., 2004).

Potentially more significant than its effects in cardioprotection is the ability of the A₃AR to inhibit tumor growth. Muscle has been found to release various small molecules, including adenosine, which have been found to inhibit cancer cell proliferation. The growth-inhibiting effects of this low-dose of adenosine was found to function through the A₃AR by decreasing the expression of protein kinase A (PKA) and subsequently increasing glycogen synthase kinase-3β (GSK-3β), β-catein and ultimately decreasing cyclin D1 and c-Myc (Fishman et al., 2000). Activation of A₃AR also decreased protein kinase B (PKB)/Akt signaling, causing a decrease in nuclear factor-κB (NF-κB) (Fishman et al., 2004). These results demonstrate A₃AR’s potential in anti-cancer or anti-metastasis therapy (Fishman et al., 2002).

The A_2AAR

It was not until the initial cloning of the A_2AAR gene (Chern et al., 1992; Fink et al., 1992) and subsequent discovery of the A_2BAR gene (Pierce et al., 1992; Stehle et al., 1992) that it was realized there were two A₂-type adenylyl cyclase-stimulating adenosine receptors. Activation of the A_2AAR causes vasodilation in the coronary artery (Belardinelli et al., 1998), and this effect can be blocked by caffeine (Belardinelli et al., 1998), a known adenosine receptor antagonist. The A_2AAR agonist CGS 21680, was found to protect the liver by attenuating ischemia-induced apoptosis when administered prior to the ischemic event (Ben-Ari et al., 2005). This protection occurs, in part, through the inhibition of the release of interferon-γ from natural killer T cells by activation of A_2AAR (Lappas et al., 2006). Additionally, protection from ishcemic events by A_2AAR appear to be mediated by activity of this receptor on bone marrow-derived cells (Day et al., 2005; Yang et al., 2005), suggesting that the modulation of the receptors on bone marrow cells could aid in recovery from ischemic events.

The most striking response to A_2AAR signaling involves the regulation of the immune response. When mice deficient in A_2AAR were challanged with endotoxin-induced sepsis, or concavavalin A-induced liver injury, the levels of pro-inflammatory cytokines greatly increased as compared to controls; in some cases, knockout mice did not recover from the sub-threshold insults (Ohta and Sitkovsky, 2001). These results suggest that the role of adenosine receptors in the downregulation of the immune response is unique and nonredundant.

A_2AAR signaling also plays a role in cancer. One theory posits that signaling downstream of the A₂-type adenosine receptors attenuates the immune response (Sitkovsky and Ohta, 2005). Accordingly, it was found that signaling through A_2AAR caused an inhibition of anti-tumor T cells, and A_2AAR-knockout mice showed a dramatic rejection of immunogenic tumors (Ohta et al., 2006). The authors state that the inactivation of anti-tumor immune cells by A_2AAR could explain why these cells are sometimes found in cancer patients, but seem to have no effect. Activation of A_2AAR has also been shown to upregulate the expression of the angiogenic factor VEGF (Ramanathan et al., 2007), which could also be involved in wound healing functions of A_2AAR signaling (Montesinos et al., 2002).

The A_2BAR

The A_2BAR gene was the most recent adenosine receptor to be identified (Pierce et al., 1992; Stehle et al., 1992), and importantly differs from other adenosine receptors in its relatively low affinity toward adenosine (Feoktistov and Biaggioni, 1997). Activation of A_2BAR was shown to stimulate growth of arterial endothelial cells (Dubey et al., 2002), yet inhibit growth of cardiac fibroblasts and aortic smooth muscle cells (Dubey et al., 1996; Dubey et al., 1998; Dubey et al., 2001). This illustrates the dichotomous activities that A_2BAR can have in different cell types. Furthermore, stimulation of A_2BAR in endothelial cells in an in vitro endothelial paracellular permeability model increased endothelial barrier function (Lennon et al., 1998). Also, during periods of increased adenosine, such as ischemia or hypoxia, A_2BAR activation reduced edema and fluid extravasation (Lennon et al., 1998). Ablation of the A_2BAR in a mouse knockout model led to an increase in leukocyte adhesion to both large and small blood vessels, and increased the size of vascular lesions post injury; this effect was mediated via bone marrow-derived cells (Yang et al., 2006; Yang et al., 2008).

The stimulation of A_2BAR was also shown to have cardioprotective effects. When stimulated after a myocardial infarction for an extended period of time, A_2BAR reduced cardiac fibrosis and remodeling, and improved overall cardiac function (Wakeno et al., 2006). CD73 is induced during ischemic preconditioning, and the generation of adenosine by CD73 is necessary for reducing infarct size. This adenosine-mediated cardioprotection is mediated via activation of A_2BAR (Eckle et al., 2007).

Inducibility of adenosine receptor genes

The basal expression of each adenosine receptor varies from cell to cell. Based on mRNA analysis of whole organs, the A₁AR and A₃AR seem to be expressed in most tissues, while expression of A_2AAR seems to be localized to lung, heart, blood vessels and immune cells. The A_2BAR is expressed in macrophages, smooth muscle cells and endothelial cells, among other cell types (Yang et al., 2006) although initially the A_2BAR was reported to be expressed at low levels in most tissues (Dixon et al., 1996). In recent years it has become apparent that in addition to having basal expression in specific cell types and within particular locations in organs, adenosine receptor expression can also be induced. Below, we will review circumstances for the induction of these receptor genes, the related mechanisms of induced expression, and some consequences of increasing the presence of adenosine receptors on a given cell.

Induction of the A₁AR gene, related mechanisms and functional consequences

The A₁AR gene structure and basal activity

The A₁AR gene was first cloned in 1991 from rat brain samples (GenBank number NM_07155) (Mahan et al., 1991). In the mouse, it consists of two exons spanning 31 kb from the translational start (ATG) to the stop codon. In the human the A₁AR gene has been reported to have up to 6 exons spaning 31 kb from the translational start to the stop codon, depending on the splice variant (Ren and Stiles, 1994a). Basal activation studies of the mouse A₁AR promoter found that transcription factors GATA4 and Nkx2.5 (see Table 1) act on sites located 500 bp upstream of the ATG, and are responsible for mouse A₁AR mRNA expression while acting synergistically (Rivkees et al., 1999). Ren and Stiles found that the 5' untranslated regions of mRNA transcripts from cells with basal (or low) level mRNA expression of the A₁AR contained different exons than transcripts found in cells with increased (or high) levels of this receptor expression, indicating that there are a variety of regulatory mechanisms at work (Ren and Stiles, 1994b).

Table 1. Transcription factors known to function on adenosine receptor gene promoters.

Location of functional transcription factor binding sites in the four adenosine receptor genes in mouse and human. The text cites the studies that describe each of the functional sites listed in this table.

		GenBank#	transcription fator site	sequence (5' to 3' on + strand)	relation to ATG
A1AR	mouse	NM_001039510	NF-κB	CTACCGGTTG	−360
			Nkx2.5	TTAAGA	−739 & −1446
			GATA4	GGGGATA	−930
	human	NM_000674	NF-κB	GGAAGTCCC	−1810
				GGGGCTCCCC	−880
A3AR	mouse	NM_009631 NM_027025	GATA6	TGCTTATCTTGATGGA	−351
			CRE	TCCAGG	−1642
	human	NM_020683 NM_001081976 NM_000677	NF-kB	GAAACCCCC	−2089
A2AAR	mouse	NM_009630	HIF-1	CACGT	−134
				GCGTG	−152
				CACGC	−1136 & −1212
			NF-κB	GAGATTCCC	−523
				GGGATCCCC	−1360
			CRE	CCTGGA	−1392 & −1412
	human	NM_000676	NF-κB	GGGAGTCTC	−580
			CRE	TCCAG	−883, −1082, & −1873
A2BAR	mouse	NM_007413	B-Myb	CAGTTG	−3243 & −4510
				CAACTG	−3276
			NF-κB	GAGAAGTCCC	−4900
				GTGAATTCCC	−5581
				GGGGCTCTC	−4033
	human	NM_000676	HIF-1	CACGTGG	−887
				CGGGAG	−873
				ACGTG	−1150

Induction of the A₁AR gene

Activation of the A₁AR under conditions that generate oxidative stress was found to have protective effects against cardiac injury (Karmazyn and Cook, 1992). Therefore, under these conditions it may be beneficial to upregulate the expression of the A₁AR receptor. To examine the role of oxidative stress in inducing the A₁AR, a smooth muscle cell line was treated with H₂O₂ or a chemotheraputic agent that generates reactive oxygen species (ROS). These treatments caused an increase in A₁AR mRNA, and this was prevented by inhibitors of the transcription fator NF-κB (Nie et al., 1998). This suggests that ROS-mediated induction of A₁AR occurs via the activation of NF-κB. Loud sound has also been found to induce ROS (Ohlemiller et al., 1999). In cochlea (inner ear) of Chinchillas exposed daily to loud noise for 6 hours, ROS was increased due to activation of NAD(P)H oxidase, which induced expression of A₁AR protein in an NF-κB-dependent manner (Ramkumar et al., 2004). Additionally, hypoxia, a condition defined by reduced oxygen availability, also generates ROS (Prabhakar and Kumar, 2004), and hypoxic conditions upregulate A₁AR protein expression in smooth muscle cells, in part by activation of NF-κB (Hammond et al., 2004).

Nitric oxide (NO) is produced endogenously at baseline and under conditions such as those involved in immune response and cellular injury (Pacher et al., 2007). NO regulation of NF-κB is dichotomous, as at high concentrations it seems to inhibit, and at low concentrations NO stimulates NF-κB activation (Umansky et al., 1998). An NO donor given to neuronal cells increases expression of A₁AR mRNA and protein, which is inhibited with an NO scavenger, and via inhibition of NF-κB (Jhaveri et al., 2006).

In vivo, the A₁AR gene has been found to be inducible in a variety of systems. In the cerebral cortex of ischemic rats the levels of A₁AR mRNA and protein were increased (Lai et al., 2005). One could suggest that the detrimental effects of ischemia could be counteracted by protective A₁AR signaling, as is seen in the cardiovascular system when A₁AR is over-expressed (Matherne et al., 1997; Dougherty et al., 1998). In the extracellular fluid of brain gliomas the concentration of adenosine is reduced by 50% (Melani et al., 2003). Interestingly, this tumor type shows an increase in A₁AR, as determined using positron emission tomography (PET) and an A₁AR-specific PET ligand (Bauer et al., 2003; Bauer et al., 2005). The relationship between the reduction in adenosine surrounding the tumors and the increase in A₁AR is intriguing and has also been demonstrated in non-tumor tissue. A mouse model which was targeted to have left ventricle dilation and dysfunction was found to have a decrease in extracellular adenosine and a concurrent increase in A₁AR (Funakoshi et al., 2007). In addition, A₁AR antagonists have been shown to increase the density of A₁ receptors and increase their function (Hettinger-Smith, 1996), while A₁AR agonists desensitize the receptors (Nie et al., 1997; Hettinger et al., 1998). These results raise the possibility that a low concentration of extracellular adenosine could signal for the increase of the A₁AR, either directly or indirectly, although the mechanisms of this hypothesis must be explored further.

Induction of the A₃AR gene, related mechanisms and functional consequences

The A₃AR gene structure and basal activity

The mouse A₃AR gene was first cloned in 1992 (Zhou et al., 1992) and two mRNA variants can be found in this species (GenBank number NM_009631 and number NM_027025). The clone described in GenBank number NM_009631 contains two exons spaning 3.1 kb from the ATG to translational stop codon. Clone number NM_027025 contains six exons and spans almost 10 kb from the ATG to the translational stop codon. The human A₃AR gene can be found in three variants, depicted in GenBank numbers NM_020683, NM_001081976, and NM_000677. The former consists of six exons and spans 19.7 kb, clone number NM_001081976 contains six exons with 80 kb from the ATG to the translational stop codon, and the later consists of two exons and spans 3.4 kb from the ATG to the translational stop codon.

A transgenic mouse was generated that used the 2.3 kb region upstream of the transcriptional start site of the mouse A₃AR gene to drive expression of the β-galactosidase (β-gal) reporter gene. Experiments with this model confirmed that the cloned gene promoter region contains all the elements required to control tissue selective expression in vivo (Yaar et al., 2002b). Putative cyclic AMP response element (CRE) and GATA putative binding sites were found in the 2.3 kb region upstream of the transcriptional start site of mouse A₃AR gene (Yaar et al., 2002a), and these are also conserved in the human A₃AR promoter (Zhao et al., 1999) (Table 1). The GATA transcription factor site was found to be important for full activation of the promoter by binding GATA6. This is interesting as GATA6 is necessary for proper cardiac embryonic development (Xin et al., 2006) and maintenance of the differentiated phenotype of smooth muscle cells (Wada et al., 2002). This suggests that the switch to a synthetic phenotype could result in changes of A₃AR expression, though these experiments have yet to be explored. Intriguingly, the CRE element in the A₃AR promoter represses expression. As described above, the A₃AR inhibits the production of cAMP. Findings by Yaar et al. show that an increase in cAMP, by direct activation of adenylyl cyclase or through adenosine activating via the A_2AAR, reduces the ability of the inhibitory transcription factor to bind to the CRE element in the promoter, leading to an increase in A₃AR expression (Yaar et al., 2002a). These results coincide with those found using an A₃AR knockout mouse and illustrate the role of A₃AR in regulating the steady-state levels of cAMP (Zhao et al., 2000).

Induction of the A₃AR gene

The anti-inflammatory effects of the A₃AR protein in regards to arthritis were identified using the selective A₃AR agonist IB-MECA in adjuvant-induced arthritis (AIA) animal models. The activation of A₃AR reduced the effects of autoimmune arthritis through suppression of tumor necrosis factor-alpha (TNF-α) (Baharav et al., 2005). As the activation of A₃AR protein can ameliorate arthritis through prevention of pro-inflammatory cytokine release, it could be a beneficial treatment to upregulate the A₃AR. This was found in two studies. In the first, the expression of A₃AR in peripheral blood mononuclear cells (PBMC) of patients with rheumatoid arthritis was compared to that of healthy patients (Madi et al., 2007); rheumatoid arthritis causes an increase in A₃AR and NF-κB proteins in PBMC. Furthermore, when PBMC isolated from healthy patients were stimulated with inflammation-inducing factors, the blockade of interleukin-2 (IL-2) or TNF-α prevented an increase in NF-κB activity and A₃AR levels, leading the authors to speculate that NF-κB aids in the induced expression of A₃AR in this inflammatory state (Madi et al., 2007). Another study sought to understand the effects of a combined treatment with methotrexate and IB-MECA. Methotrexate is a drug used to treat inflammation caused by rheumatoid arthritis (Weinblatt et al., 1985). It mediates its effects by inhibiting the activity of adenosine deaminase (ADA) and AMP deaminase. The eventual outcome is to increase the concentration of adenosine (Baggott et al., 1986). Methotrexate increased the expression of A₃AR mRNA and protein levels, and the addition of ADA reversed the effects of methotrexate (Ochaion et al., 2006).

Breast and colon carcinoma cells have increased levels of A₃AR mRNA and protein compared to adjacent non-cancerous tissue (Madi et al., 2004). The physiological levels of adenosine do not reach the concentrations needed to activate A₃AR (Schutle and Fredholm, 2002). It was speculated that more A₃AR expression by the tumor cells may be triggered by the elevation in the extracellular adenosine released by necrotic or hypoxic cells in the microenvironment of the tumor. In addition, upregulation of transcription factors, such as NF-κB, in tumor cells may be responsible for A₃AR upregulation. The exact mechanism of how a cancerous cells upregulate the A₃AR gene has yet to be determined. The upregulation of the A₃AR could be developed as a cancer detection tool, as an increase in A₃AR on peripheral blood cells was found in patients with colon cancer (Gessi et al., 2004). Moreover, stimulation of A₃AR on cancer cells has been suggested to have the potential to become a cancer therapy (Fishman et al., 2002).

Induction of the A_2AAR gene, related mechanisms and functional consequences

The A_2AAR gene structure and basal activity

The A_2AAR gene was first cloned from mouse in 1989 (GenBank number NM_009630) (Libert et al., 1989). The mouse A_2AAR gene consists of two coding exons and spans 7.8 kb from the ATG to the translational stop site. The human A_2AAR gene (GenBank number NM_000675) contains two coding exons and measures 8.1 kb from the ATG to the translational stop codon. Studies with the rat A_2AAR gene (NM_053294) found that the transcription factor nuclear factor (NF)-1 acts as an inhibitor of A_2AAR gene expression in a cell-specific fashion (Lee et al., 2003b). NF-1 has been described to act as both a postive and negative regulator of gene transcription (Gao and Kunos, 1998), where it cooperates with other factors that direct it to activate or inhibit transcription (Matsumoto et al., 1989; Roy and Guerin, 1993).

Induction of the A_2AAR gene

The first studies regarding the induciblity of the A_2AAR gene were conducted soon after it’s cloning. A_2AAR mRNA expression was found to be negatively regulated by its own activation via the generation of cAMP (Saitoh et al., 1994). Contradicting this finding, a recent study showed that the mutated gene in patients with Huntington’s disease was found to negatively suppress A_2AAR expression (Chiang et al., 2005). The trinucleotide repeat in the mutated gene blocks the binding of a cyclic AMP response element-binding protein (CREB) to the core promoter structure; an element that positively regulates A_2AAR expression in the normal gene (Chiang et al., 2005). Taken together, these studies suggest a complex mechanism involved in mediating the expression of A_2AAR by cAMP.

A_2AAR protein activation aids in the regulation of the immune response (Sitkovsky and Ohta, 2005), so it is not surprising that this receptor is also upregulated by inflammatory cytokines. The treatment of mouse macrophages with bacterial lipopolysaccharide (LPS) induces the expression of A_2AAR mRNA and the pro-inflammatory cytokine TNF-α (Murphree et al., 2005). Both A_2AAR and TNF-α expression by LPS are inhibited by BAY 11-7082, an NF-κB inhibitor, indicating that this transcription factor is important for A_2AAR induciblity (Murphree et al., 2005). When LPS-treated cells were given an A_2AAR-specific agonist, TNF-α release was decreased, suggesting that the immune response initiated by LPS can be attenuated via feed-back through A_2AAR (Murphree et al., 2005). This effect was also seen when comparing peripheral blood mononuclear (PBM) cells of patients with and without congestive heart failure (CHF) (Capecchi et al., 2005). In the CHF samples, A_2AAR mRNA and protein were found to be upregulated compared to healthy controls, and this upregulation resulted in a decrease in TNF-α in the plasma of CHF patients (Capecchi et al., 2005). Lung epithelial cells showed a similar upregulation of A_2AAR in response to TNF-α stimulation as well as interleukin-1β (IL-1β), and the mechanism of this upregulation involves NF-κB (Morello et al., 2006).

Heme oxygenase-1 (HO-1) is an enzyme that breaks down heme molecules into iron, biliverdin and carbon monxide (CO), and this catabolism increases ATP and adenosine (Tsui et al., 2005). HO-1 is now recognized for its involvement in modulating anti-oxidant and anti-inflammatory activities (Korthuis and Durante, 2005). A recent study found that adenosine signaling via A_2AAR induced HO-1 expression, and that CO generated from HO-1 activity subsequently upregulated A_2AAR mRNA and protein (Haschemi et al., 2007). The authors of this study suggest a potential mechanism of A_2AAR upregulation involving the ability of CO to increase ROS in a similar fashion as hypoxia (Bilban et al., 2006; D'Amico et al., 2006; Nakahira et al., 2006). Preliminary, unpublished studies from their lab (Haschemi et al., 2007) indicate that hypoxia-inducible factor -1 (HIF-1), activated under conditions of hypoxia or oxidative stress (Schofield and Ratcliffe, 2004; Schofield and Ratcliffe, 2005) is also activated by CO, suggesting that this transcription factor could be involved in A_2AAR mRNA induction. Indeed, A_2AAR mRNA and protein levels are increased under hypoxic conditions (Kobayashi et al., 1998). The potential involvement of HIF-1 in the upregulation of A_2AAR is provocative, as A_2AAR stimulation was recently found to lead to the activation HIF-1 (De Ponti et al., 2007) (see Table 1).

The mechanism of A_2AAR upregulation in the brain is less understood. Using a transgenic mouse with a portion of the A_2AAR promoter fused to the gene encoding β-galactosidase, A_2AAR was found to be highly expressed in localized regions of the brain (Lee et al., 2003a). Subsequent studies have sought to understand the regulatory mechanisms involved in diseases such as Parkinson’s and schizophrenia (Fuxe et al., 2005). A_2A–dopamine D₂ heteromeric receptor complexes constitutively exist in striatal tissue and A_2A receptors antagonize dopamine D₂ receptor agonist recognition and signaling and modulate D₂ receptor trafficking. Additionally, A_2A receptor antagonists increase the therapeutic index of L-DOPA (Levadopa or 3,4-dihydroxy-L-phenylalanine) and D₂ receptor agonists. Treatments of schizophrenia may be A_2A receptor agonists that reduce D₂ receptor signaling in A_2A–D₂ receptor heteromers (Fuxe et al., 2005). A side effect of the Parkinson’s drug L-DOPA is involuntary movements (referred to as L-DOPA-induced dyskinesia). L-DOPA also induces A_2AAR mRNA upregulation (Tomiyama et al., 2004), and subsequent activation of A_2AAR has been shown to reduce dyskinesia in response to L-DOPA (Agnati et al., 2004; Tronci et al., 2007).

Induction of the A_2BAR gene, related mechanisms and functional consequences

The A_2BAR gene structure

The A_2BAR gene was first cloned from human (GenBank number NM_000676) and rat (GenBank number NM_017161) (Pierce et al., 1992; Stehle et al., 1992). The human A_2BAR gene spans 30 kb from the ATG to the translational stop codon and contains two exons. The mouse A_2BAR gene (GenBank number NM_007413) contains two coding exons in 16.6 kb of genomic DNA from the ATG to the translational stop codon (Stehle et al., 1992).

Induction of the A_2BAR gene

A novel A_2BAR knockout/β-gal knockin mouse was generated such that the endogenous A_2BAR promoter drove expression of the bacterial β-gal gene. This created a tool whereby one can visualize individual cells and tissues where this promoter is active, as opposed to having to rely on the crude extraction of mRNA from tissues containing multiple cell types (Yang et al., 2006). Using this mouse, A_2BAR expression was found to be localized in the vasculature of most organs, including patchy expression in the smooth muscle and endothelial cells, as well as macrophages (Yang et al., 2006). Additional studies with vascular smooth muscle cells obtained from this knockout, as well as studies with A_2BAR promoter constructs, confirmed that A_2BAR in mice is positively regulated by the proliferation-dependent transcription factor B-Myb (St. Hilaire et al., 2008). These findings coincide with studies that aim to understand factors involved in the ability of the transcription factor v-Myb to transform cells. The transcription factor v-Myb was found to upregulate the A_2BAR gene in chicken (Worpenberg et al., 1997; Kattmann and Klempnauer, 2002). Previous studies indicated that A_2BAR activation inhibits proliferation in human and rat aortic smooth muscle cells (Dubey et al., 1996; Dubey et al., 1998). These regulatory studies suggest that A_2BAR could be involved in a feed-back mechanism, where proliferation induces A_2BAR expression and subsequent activation of A_2BAR can inhibit the cell growth. This concept is of interest in the context of cardiovascular disease, where excessive vascular smooth muscle cell proliferation is a hallmark (Hao et al., 2003); agonists of A_2BAR could have a potential role in halting this proliferation.

Like A_2AAR, A_2BAR expression is also regulated by associated conditions such as inflammation and hypoxia and these stress conditions generate extracellular adenosine (Sitkovsky and Ohta, 2005). Interferon-γ (IFN-γ) is a cytokine that binds tyrosine kinase receptors and activates signaling pathways involving JAK and STAT transcription factors (van Boxel-Dezaire and Stark, 2007). IFN-γ signaling upregulates A_2BAR mRNA in mouse macrophage cells, and subsequent activation with A_2BAR-agonists reduces IFN-γ-induced expression of cytokines, NO synthase, and major histocompatibility complex (MHC) class II genes, suggesting that A_2BAR is involved in a feed-back mechanism for deactivating macrophages (Xaus et al., 1999). It is interesting to note that in intestinal epithelial cells, IFN-γ was found to decrease A_2BAR signaling by downregulation of adenylyl cyclase (Kolachala et al., 2005b), suggesting that the same signal coupled to A_2BAR activity is differentially regulated in distinct cell types.

TNF-α is a cytokine that is involved in a variety of inflammatory responses and stressful conditions such as hypoxia and ischemia, and upregulates the transcription of additional cytokines involved in inflammation (Popa et al., 2007). In human astrocyte cells, TNF-α was found to increase selectively the receptor activity of A_2BAR without increasing mRNA or protein levels, possibly by inhibiting receptor desensitization (Trincavelli et al., 2004). The other G_s coupled adenosine receptor, the A_2AAR, which has been shown to be upregulated by TNF-α in other cell types (Morello et al., 2006), was not affected by TNF-α in this population of human astroglial cells. Interestingly, in a colitis model using human colonic epithelia, A_2BAR was found to be upregulated by TNF-α signaling (Kolachala et al., 2005a). Accordingly, unpublished work by our group using vascular smooth muscle cells indicates that TNF-α upregulates A_2BAR expression, potentially via the activation of NF-κB and/or HIF-1 sites within the promoter.

Hypoxia is a state of low oxygen that can be found in such conditions as cardiovascular ischemia and tumors (Hirota and Semenza, 2006). One mechanism to reverse hypoxic conditions is to increase the blood supply to the area. Angiogenesis is a process characterized by the growth of new blood vessels, and adenosine is one signal that initiates this process (Sitkovsky et al., 2004). Lung fibroblast cells exposed to hypoxic conditions show increased expression of A_2BAR mRNA (Zhong et al., 2005). The activation of A_2BAR was further shown to induce expression of angiogenic factors IL-8 and VEGF in endothelial cells (Feoktistov et al., 2002) and mast cells (Feoktistov et al., 2003). HIF-1 is activated and translocated to the nucleus under hypoxic conditions leading to the upregulation of genes, such as erythropoietin (Semenza et al., 1991) and VEGF (Liu et al., 1995). Under hypoxic conditions, HIF-1 acts on the A_2BAR gene promoter in human endothelial cells, increasing A_2BAR mRNA and protein levels (Kong et al., 2006). The upregulation of A_2BAR protein increases the barrier function of the endothelial cells and increases angiogenesis (Kong et al., 2006).

An overview of the above reported functions and regulatory factors of adenosine receptors is summerized in Table 3.

Table 3. Reported functions and regulatory factors of adenosine receptors.

Adenosine receptor	Main Functions	Regulatory factors of receptor level or/and activity
A1AR	Vasculature protection; Angiogenesis promotion; Anticonvulsant effects	Reactive oxygen species (ROS); Nitric oxide (NO); Ischemia; Adenosine
A3AR	Cardioprotective effects; Anti-inflammatory effects; Tumor growth prevention	cAMP; Inflammation; Cancer
A2AAR	Vasodilation of the coronary artery; Liver protection; Downregulation of immune response; Upregulation of VEGF in macrophages; Wound healing promotion	cAMP; Inflammatory cytokines; Carbon monoxide (CO); Hypoxia; L-DOPA
A2BAR	Growth stimulation of arterial endothelial cells; Growth inhibition of cardiac fibroblasts and aortic smooth muscle cells; Increased endothelial barrier function; Reduced edema and fluid extravasations; Modulation of inflammatory cytokines; Leukocyte adhesion and rolling on blood vessels; Vascular protection; Cardioprotective effects	Proliferation; Inflammation; Hypoxia

Putative transcription factor binding sites in adenosine receptor genes: a potential predictor of functional responses

Analysis of the presence of confirmed and putative transcription factor binding sites within adenosine receptors reveals that many of these sites are shared among promoters (Table 1). Table 2 depicts putative transcription factor binding sites found in the gene promoters of the adenosine receptors. Transcription factor binding sites for Nkx2.5 are found in the promoters of all the adenosine receptors in mouse and human; studies mentioned above have found these sites functional in the mouse A₁AR. Nkx2.5 is a homeobox gene highly expressed in the developing heart of vertebrates. Nxk2.5 drives expression of genes such as cardiac α-actin, myocardin, endothelin-converting enzyme-1, and the A₁AR (Akazawa and Komuro, 2005). Nkx2.5 is also expressed in the adult heart and was recently reported to function in the maintenance of the fully differentiated cardiac phenotype (Toko et al., 2002). All four of the adenosine receptors have been shown to be expressed in, and have protective effects on the cardiovasculature. As putative Nkx2.5 sites are found within the promoters of all four adenosine receptors, one can speculate on the potential of Nxk2.5 as a global regulator of the adenosine receptors in the heart and vasculature. Nkx2.5 and GATA4 were found to function as cofactors, which is significant as putative sites for these two transcription factors are located in the promoters of mouse A₁AR and A_2AAR. GATA4, 5 and 6 are expressed in developing hearts (Charron and Nemer, 1999). GATA6 was found to be involved in the upregulation of the mouse A₃AR promoter, and putative GATA6 sites are found in the human A₃AR and the mouse A_2BAR. However, GATA6 cannot be substituted for GATA4 as an Nkx2.5 cofactor (Durocher et al., 1997).

Table 2. Putative binding sites in adenosine receptor gene promoters.

The promoters of the four adenosine receptor genes in mouse and human were searched for putative transcription factor binding sites using MacVector 9.0 nucleic acids subsequences search. Promoters were searched for transcription factors that are known to have activity on at least one other adenosine receptor (see Table 1).

		GenBank#	transcription fator site	sequence (5' to 3' on + strand)	location in relation to ATG
A1AR	mouse	NM_001039510	NF-kB	CTACCGGTTG	−390
			Nkx2.5	TTAAGA	−793, −1456
			GATA4	GGGGATA	−930
	human	NM_000674	CRE	TCCAGG	−124, −765
			CRE	CCTGGA	−132, −459, −1005
			B-Myb	CAGTTG	−2728
			B-Myb	CAACTG	−2259
			HIF-1	GCGTG	−341
			HIF-1	CACGT	−660
			HIF-1	ACGTG	−2169
			HIF-1	CACGC	−717
A3AR	mouse	NM_009631 NM_027025	Nkx2.5	TTCTTAA	−1078
			CRE	CCTGGA	−1392, −1412, −2124
			CRE	TCCAGG	−3135
			GATA4	GTATCCC	−3298
	human	NM_020683 NM_001081976 NM_000677	Nkx2.5 Nkx2.5 B-Myb HIF-1 HIF-1	TAAAGTG CACTTGA CAGTTG GCGTG ACGTG	−509 −2005 −4504 −1280, −2670, −4109 −1802, −2720
A2AAR	mouse	NM_009630	HIF-1	CACGC	−158, −1091, −1176, −2464, −4588
			HIF-1	GCGTG	−1102, −1589
			CRE	CGTCA	−398, −3675
			CRE	TCCAGG	−1010
			Nkx2.5	CACTTGA	−2225, −4034
			GATA4	GGATAC	−4721
	human	NM_000676	CRE	CCTGGA	−284, −506, −632
			CRE	CCTGGA	−343
			B-Myb	CAACTG	−2099 & −2198
			NF-kB	GGGGTTTTTC	−1930
			Nkx2.5	TCAAGTG	−1565
A2BAR	mouse	NM_007413	B-Myb	CAGTTG	−832
			Nkx2.5	TAAAGTG & TGAAGTG	−1021 & −1107
			HIF-1	CCCACGTA	−1295
			NF-kB	GGGAATTTC	−2228
	human	NM_000676	GATA6	TTATCTT	−369
			CRE	CGTCA	−723
			HIF-1	CACGT, GCGTG, CACGC	−987, −1138, −1212
			Nkx2.5	TGAAGTG	−1365–2203
			GATA4	GGATAC	−1477
			B-Myb	CAGTTG	−1583

HIF-1, was found to function in the expression of the mouse A_2AAR and the human A_2BAR genes, and putative HIF-1 sites are found in all of the adenosine receptor promoters in mouse and human (Schofield and Ratcliffe, 2004; Lopez-Lazaro, 2006). A recent review by Sitkovsky and Lukashev highlights the interplay between hypoxia, the immune response and the A₂-adenosine receptors. The authors of this review hypothesize that the secondary damage to the microvasculature at sites of inflammation results in hypoxia, which serves as a signal to attenuate the immune response by increasing the amounts of A₂AR in immune cells (via HIF-1), with parallel inhibition of adenosine kinase and upregulation of CD73 (Sitkovsky and Lukashev, 2005). Signaling via the A_2AAR and A_2BAR on these cells inhibits the production of pro-inflammatory cytokines such as IFN-γ (Xaus et al., 1999; Lappas et al., 2005). Future studies should explore HIF-1 activities on the promoters of the adenylyl cyclase-inhibiting A₁AR and A₃AR. In immune cells, a balance may be achieved as to the amount of the different adenosine receptors expressed that will determine if that cell proceeds with a pro- or anti-inflammatory response.

The transcription factor NF-κB is activated by a variety of inflammatory stimuli and mediates the upregulation of genes involved in the immune response. NF-κB increases expression of the mouse A₁AR, A_2AAR and A_2BAR, and the human A₁AR, A₃AR and A_2AAR, mRNA and/or protein levels. Putative NF-κB sites are also found in the mouse A₃AR and the human A_2BAR promoters. Studies are needed on the concerted upregulation of the adenosine receptors by NF-κB. As mentioned in regards to HIF-1, we must try to understand the effects of adenylyl cyclase-stimulating and inhibiting adenosine receptors being activated simultaneously on the same cell.

CRE were found to be functional in the mouse A₃AR and human A_2AAR gene promoters, and putative sites are found in all other adenosine receptor promoters. As adenosine receptor activation either inhibits (A₁AR and A₃AR) or increases (A_2AAR and A_2BAR) the generation of cAMP by adenylyl cyclase, it is intriguing that all adenosine receptors have putative CRE sites in their promoters, causing one to speculate whether their activation also involves auto-regulatory processes, such as those mentioned above in the case of the mouse A₃AR.

Significance of understanding the regulation of adenosine receptors

Expression of all four of the adenosine receptors can be induced. As they are involved in, and regulate various mechanisms – e.g. cell growth and proliferation, apoptosis, immune response, and angiogenesis – the ability to modulate the expression and activity of these receptors in a disease state could be beneficial and should be explored as potential therapies in diseases such as asthma, Parkinson’s disease and cancer. Similarly, adenosine receptors modulate feed-back-type responses, indicating their utility as regulators of cellular processes responsible for their activation. It should be emphasized, however, that upregulated adenosine receptor gene expression does not always result in elevated levels of active receptor. The latter depends on the proper localization and coupling of the receptor, among other factors. Future studies should continue to examine parallel changes (or lack of) in adenosine receptor gene expression and receptor activity under various conditions.