Purinergic regulation of epithelial transport

Abstract

Purinergic receptors are a family of ubiquitous transmembrane receptors comprising two classes, P1 and P2 receptors, which are activated by adenosine and extracellular nucleotides (i.e. ATP, ADP, UTP and UDP), respectively. These receptors play a significant role in regulating ion transport in epithelial tissues through a variety of intracellular signalling pathways. Activation of these receptors is partially dependent on ATP (or UTP) release from cells and its subsequent metabolism, and this release can be triggered by a number of stimuli, often in the setting of cellular damage. The function of P2Y receptor stimulation is primarily via signalling through the G_q/PLC-β pathway and subsequent activation of Ca²⁺-dependent ion channels. P1 signalling is complex, with each of the four P1 receptors A₁, A_2A, A_2B, and A₃ having a unique role in different epithelial tissue types. In colonic epithelium the A_2B receptor plays a prominent role in regulating Cl⁻ and water secretion. In airway epithelium, A_2B and A₁ receptors are implicated in the control of Cl⁻ and other currents. In the renal tubular epithelium, A₁, A_2A, and A₃ receptors have all been identified as playing a role in controlling the ionic composition of the lumenal fluid. Here we discuss the intracellular signalling pathways for each of these receptors in various epithelial tissues and their roles in pathophysiological conditions such as cystic fibrosis.

Extracellular purines such as ATP and adenosine signal through membrane-associated purinergic receptors as autocrine and paracrine substances as well as neurotransmitters. The ubiquitous nature of these ligands as well as the abundance of expression of purinergic receptors account for purinergic control of diverse effects in many tissue types. Control of ion and fluid transport across epithelia is one such system in which purinergic regulation is highly significant, even considering the variety of hormones and other regulators that are well known to regulate epithelial transport. Purinergic receptors have been identified in most epithelial tissues and have been well characterized, specifically in the gastrointestinal, airway and kidney epithelia.

This large family of purinergic receptors has been subdivided into two major classes, P1 and P2, that have preferential affinity for adenosine and ATP, respectively. Figure 1 gives an overview of these receptors and their role in regulation of ion transport. The P1 receptors are a family of G protein-coupled receptors that signal through multiple intracellular effectors in response to nucleoside activation, primarily with adenosine. There are four known subtypes of P1 receptors, A₁, A_2A, A_2B, and A₃, each of which has been cloned and characterized biochemically and pharmacologically. The A₁ receptor signals through the G_i/o family of G proteins, resulting most significantly in an inhibition of adenylyl cyclase but also, via βγ subunits, in the activation of phospholipase C-β (PLC-β), the activation of several K⁺ channels, and the inactivation of Ca²⁺ channels, among others. The A_2A and A_2B receptors classically signal primarily via G_s, resulting in activation of adenylyl cyclase, an increase in cAMP production, and subsequent activation of protein kinase A (PKA). In some cases activation of the A_2B receptor has been documented to signal through other effectors, specifically mobilization of intracellular calcium, and this may result from interaction with G_q or other G proteins (Feoktistov & Biaggioni, 1997; Linden et al. 1999). It is also possible that in some cells cAMP may mobilize Ca²⁺ due to activation of phospho lipase C-ε (PLC-ε) via EPAC (exchange protein activated by cAMP) and RAP2B (Evellin et al. 2002). The A₃ receptor signals primarily via G_i/o proteins. The primary downstream effects of A₃ activation are inhibition of adenylyl cyclase function and, in some cases, activation of phospholipase C to provide an intracellular calcium signal. Inosine, a purine nucleoside metabolite of adenosine, has activity at the A₃ receptor but has weak activity at the other P1 receptors (Jin et al. 1997; Ralevic & Burnstock, 1998; Fredholm et al. 2001).

Figure 1. Purinergic receptors important in regulating epithelial ion transport.

The P1 receptors A_2A and A_2B classically signal through G_s, resulting in an increase in cAMP and activation of PKA. A rise in intracellular Ca²⁺ noted in some cells in response to A_2B receptor activation may result from receptor coupling to G_q and activation of PLC-β or from cAMP-activation of Rap2B that raises intracellular Ca²⁺ via PLC-ɛ. This rise in intracellular Ca²⁺ by either of these mechanisms likely accounts for the activation of PLA₂ that contributes to the A_2B activated Cl⁻ current. The A₁ receptor has been shown to be active in airway epithelial cell lines and evokes a Ca²⁺ response through the βγ subunits of G_i. This triggers basolateral K⁺ efflux and may trigger Cl⁻ release. P2X receptors are ligand-gated ion channels and increase cell permeability to Na⁺ or Ca²⁺. P2Y receptors classically signal through G_q, resulting in an increase in intracellular Ca²⁺ and activation of Ca²⁺-activated Cl⁻ and K⁺ channels, although a Ca²⁺-independent Cl⁻ current resulting from P2Y activation has been identified.

The P2 family of receptors is further divided into metabotropic (G protein-coupled) P2Y receptors and ionotropic P2X receptors. There are currently eight known P2Y receptors (Y₁, Y₂, Y₄, Y₆, Y₁₁, Y₁₂, Y₁₃ and Y₁₄) and eight P2X receptors (P2X_1-7 and P2XM). These receptors are classically activated by purine nucleotides, and the differing affinities for specific diphosphate or triphosphate nucleotides help to differentiate among similar receptor subtypes (Fredholm et al. 1994; Harden et al. 1997). The P2Y receptor preferences for naturally occurring ligands are summarized in Table 1 (King et al. 2001; Communi et al. 2001). A subset of these receptors can be activated by the pyrimidine nucleotides UTP and UDP, despite their original classification as ‘purinergic’ receptors. A receptor formally known as the UDP-glucose receptor has recently been added to the P2Y family and designated P2Y₁₄ (Abbracchio et al. 2003). P2Y receptors classically signal through a G_q-dependent pathway, activating phospholipase C and mobilizing intracellular Ca²⁺, but some can modulate other effector pathways as well. P2X receptors are all activated by ATP, and, excepting P2X₆, all are also activated by 2-MeSATP. BzATP is a selective agonist of P2X₅ and P2X₇ receptors (Khakh et al. 2001; Bo et al. 2003). P2X receptors are ligand-gated channels and their activation increases the permeability of the plasma membrane specifically to Na⁺ or Ca²⁺. In some cases antagonists have been developed which discriminate between P2X and P2Y receptor subtypes. These include the selective P2Y₁ antagonist, MRS2279 (Boyer et al. 2002), and the selective P2Y₁₂ antagonist AR-C69931 (Simon et al. 2002).

Table 1.

Natural ligands for P2Y receptors

Receptor	Preferred natural agonists
P2Y1	ADP > ATP
P2Y2	UTP = ATP
P2Y4	UTP ≫ ATP
P2Y6	UDP > UTP > ADP
P2Y11	ATP
P2Y12	ADP > ATP
P2Y13	ADP > ATP
P2Y14	UDP-glucose > UDP-galactose > UDP-N-acetylglucosamine

In order to initiate signalling via purinergic receptors, nucleotides and their metabolites, specifically ATP, UTP, ADP, UDP, UDP-glucose or adenosine, must be present in the extracellular fluid at concentrations sufficient to activate receptors. Within cells, ATP levels are high and maintained within a narrow range. A variety of stimuli are capable of initiating ATP (or UTP) release from intact cells including mechanical stimuli, cell swelling, and inflammatory stimuli (Taylor et al. 1998; Roman & Fitz, 1999; Harden & Lazarowski, 1999). Once released, ATP at the luminal surface is able to activate P2 receptors or be readily converted to ADP and AMP via widely expressed ecto-nucleotidases present on the surface of cells (Zimmermann, 1996). AMP is converted to adenosine via the action of ecto-5′nucleotidase. Adenosine can be taken up into cells by nucleoside transporters and then be reconverted to AMP by adenosine kinase. Nucleoside transporters are thought to play a significant role in governing the effect of adenosine at the epithelial surface by controlling the extracellular concentration (Szkotak et al. 2001, 2003). Adenosine can be converted to inosine by the action of adenosine deaminase, found both inside and on the surface of cells. This ready release of nucleotides with a variety of insults and their subsequent metabolism have coevolved with multiple extracellular nucleotide signalling pathways that can utilize both ATP (or UTP) and adenosine and can efficiently work together to regulate ion transport across the epithelial surface. One commonly held belief is that, in many tissue types, this stimulation of ion transport and the flux of water that accompanies it are a natural defense system that functions to effectively wash away noxious stimuli in the setting of cellular damage or inflammation (Lazarowski & Boucher, 2001; Leipziger, 2003).

P2 receptors in epithelial transport

The role of P2 receptors in regulating epithelial transport has been recently reviewed by Leipziger (2003), and for that reason the majority of this review will focus instead on the role of P1 receptors in regulating secretion across the epithelium. It is worthwhile, however, discussing the basic signalling via P2 receptors in major tissue epithelia since extracellular adenosine is thought to arise from the breakdown of extracellular ATP, and therefore P1 and P2 signalling events are often triggered simultaneously by concurrent increases in extracellular ATP and adenosine.

P2 receptor signalling in gastrointestinal epithelia

P2 signalling is important in the gastrointestinal tract, where, in summary, P2 activation leads to activation of K⁺, HCO₃⁻, and Cl⁻ secretion and inhibition of Na⁺ reabsorption at various locations in the gallbladder, pancreas, small intestine, and colon (Roman & Fitz, 1999; Leipziger, 2003). There are a variety of P2 receptors implicated in regulating transport along the gastrointestinal tract, including some that have been well characterized: the P2Y₆ receptor in the mouse gallbladder, the P2Y₄ receptor in mouse jejunum, the P2Y₁ and P2Y₆ receptors in rat distal colon, the P2Y₄ receptor in rat pancreas, and the P2Y₂ receptor in guinea-pig pancreas (Leipziger et al. 1997; Cressman et al. 1999; Ishiguro et al. 1999; Hede et al. 1999; Kottgen et al. 2003). In gastrointestinal epithelial tissue, ATP does not directly activate Cl⁻ secretion via P2 receptors since the cAMP-dependent cystic fibrosis transmembrane conductance regulator (CFTR) Cl⁻ channel is thought to be the only chloride channel expressed, although there is longstanding debate on the existence of a Ca²⁺-dependent Cl⁻ channel in the colonic epithelium (Anderson et al. 1992; Wagner et al. 1992; Rozmahel et al. 1996; Grubb & Gabriel, 1997; Barrett & Keely, 2000). Much of the controversy around this topic originates from the use of colonic epithelial cell lines such as T84 cells that express Ca²⁺-activated Cl⁻ channels in their undifferentiated state, but down-regulate them upon forming differentiated monolayers (Anderson & Welsh, 1991; Morris et al. 1992). Despite the presumed absence of a Ca²⁺-activated Cl⁻ channel in differentiated cell lines and in in vivo epithelial tissue, activation of P2 receptors can still lead to lumenal fluid accumulation via stimulation of K⁺ secretion and inhibition of Na⁺ reabsorption (Leipziger, 2003). Thus, the action of ATP in the colon is thought to stimulate water efflux as a protective response, as in other secretory epithelia. In the case of the colon, this would present clinically as secretory diarrhoea in response to an inflammatory stimulus.

P2 receptor signalling in respiratory epithelia

In bronchial epithelium, signalling via P2 receptors has received much attention because of the possible therapeutic potential in treating cystic fibrosis. The predominant P2 receptor in respiratory epithelium is the P2Y₂ receptor that classically couples to G_q, increases intracellular Ca²⁺, and activates Ca²⁺-dependent Cl⁻ channels, although other pathways for Cl⁻ secretion in airways have been identified. Specifically, a Ca²⁺-independent and adenosine receptor independent Cl⁻ current can be triggered by ATP or UTP in airway epithelia (Stutts et al. 1994; Inglis et al. 1999). The importance of the P2Y₂ receptor in the mouse tracheal epithelium was confirmed in a knockout study in which mice deficient in this receptor lost 85–95% of the nucleotide-stimulated Cl⁻ secretion in this tissue (Cressman et al. 1999). The net effect of P2Y₂ activation is to increase Cl⁻ secretion and K⁺ secretion and inhibit electrogenic Na⁺ absorption, all of which lead to water secretion (Mason et al. 1991; Clarke et al. 1997; Inglis et al. 1999; Leipziger, 2003). Because this pathway involves Ca²⁺-dependent Cl⁻ channels rather than the cAMP-dependent cystic fibrosis transmembrane conductance regulator (CFTR) Cl⁻ channel, activation of the P2Y₂ receptor has been explored as a potential method of overcoming the Cl⁻ secretion defect in cystic fibrosis.

P2 receptor signalling in renal epithelia

P2 signalling along the renal epithelium is similar to that found in the respiratory epithelium and gastrointestinal epithelium in that P2 receptor activation causes a rise in intracellular Ca²⁺ that initiates ion transport. The renal epithelium is complicated by the existence of multiple receptor subtypes and individual variations within each segment of the nephron. The role of each receptor subtype remains to be elucidated. In the nephron of both the rat and mouse, it is currently thought that P2Y₁ and P2Y₂ receptors play a primary role in epithelial tissues along the nephron with the P2Y₆ receptor also recently implicated in the rat proximal collecting duct, thick ascending limb, and thin descending limb (Bailey et al. 2000; Schwiebert & Kishore, 2001). These receptors are of interest in this tissue primarily because of their role in regulating salt and water balance. Salt and water reabsorption are inhibited by activation of P2 receptors along the tubular lumen, making manipulation of these receptors of interest in hypertension (Kishore et al. 1995; McCoy et al. 1999; Leipziger, 2003). It has also been shown that P2 receptors play a significant role in regulating water balance independently of vasopressin. Hypervolaemia and ischaemic reperfusion injury are known to induce P2 expression along the nephron, while hypovolaemia down-regulates P2 expression, implicating these receptors in diuresis or water retention in certain pathophysiological states (Kishore et al. 1998, 2000; Schwiebert & Kishore, 2001).

P1 receptors in epithelial transport

Adenosine-mediated control of ion transport has been demonstrated in a variety of epithelial tissues and as a result of activation of all four P1 receptor subtypes. P1-receptor-mediated control of Cl⁻ secretion via the cAMP-activated CFTR Cl⁻ channel is among the most significant and most studied effects of adenosine on epithelial cells. Ca²⁺-activated currents stimulated primarily via the G_i/PLC-β pathway also play a major role in the diverse effects of P1 receptors in these tissues.

P1 receptor signalling in gastrointestinal epithelia

The best characterized tissue with regard to P1-mediated ion transport is the colonic epithelium. The interest in this tissue arises from the potent Cl⁻ and water secretion stimulated by adenosine via the A_2B receptor, which is clinically relevant as a contributing factor in secretory diarrhoea in the setting of inflammation and also as a possible target for treatment of cystic fibrosis. It is important to note that adenosine, unlike ATP acting via a P2 receptor, is known to have a direct effect on Cl⁻ release in the colon via activation of CFTR. The A_2B receptor has been identified on both the mucosal and basolateral aspects of colonic epithelial cells, and activation at either site results in Cl⁻ secretion (Barrett et al. 1989; Strohmeier et al. 1995). A similar effect of A_2B stimulation has been described in isolated rabbit ileum (Dobbins et al. 1984). A major source of adenosine in the bacteria-infected gut is neutrophils, which release adenine nucleotides that are rapidly converted to AMP and then to adenosine by 5′-nucleotidase, which is highly expressed on gut epithelium (Madara et al. 1993). The classical pathway for A_2B-mediated Cl⁻ release is via G_s activation of adenylyl cyclase, with a rise in cAMP that directly activates the CFTR Cl⁻ channel. It has long been speculated, however, that another pathway may be involved based upon the saturation of Cl⁻ secretion at levels of A_2B stimulation that fail to saturate cAMP levels and the ability of adenosine to stimulate a Cl⁻ efflux equal to that of forskolin at much lower cAMP levels (Barrett et al. 1989, 1990; Strohmeier et al. 1995; Clancy et al. 1999; Huang et al. 2001). This effect has been attributed to increased coupling of the A_2B receptor to CFTR as well as to the existence of an alternative second messenger, and it now seems that both of these are likely to play a role in A_2B-mediated Cl⁻ secretion. With respect to coupling, A_2B adenosine receptors are known to couple tightly to CFTR as well as adenylyl cyclase and PKA in airway epithelial cell models, as will be discussed later (Huang et al. 2001). With respect to an additional second messenger, a pathway that has recently been implicated in A_2B-mediated Cl⁻ secretion is the activation of phospholipase A₂ (PLA₂) in parallel with the activation of adenylyl cyclase (Barrett & Bigby, 1993; Bouritius et al. 1999; Cobb et al. 2002). It has been proposed that activation of PLA₂ contributes to Cl⁻ secretion via the effects of arachadonic acid (AA) or its metabolites, the precise mechanism of action of which is unknown but may be through effects on the electrochemical driving force for ion transport through CFTR (i.e. by activation of K⁺ channels), through direct effects of lipid species on CFTR or associated factors, or by locally increasing cAMP production in the vicinity of CFTR, an effect which may not be recognized with total cellular cAMP assays (Cobb et al. 2002).

P1 receptor signalling in airway epithelia

A_2B activation of Cl⁻ secretion at the lung epithelium is similar to that found in the colon, in which an increase in apical membrane Cl⁻ conductance is produced through a G_s pathway activating adenylyl cyclase and raising cAMP (Lazarowski et al. 1992; Huang et al. 2001). This has been found in a transformed cell line from normal human airway epithelia (BEAS39), in a model for polarized serous cells thought to play a critical role in liquid secretion in cystic fibrosis (Calu-3), and in primary cultures of human nasal epithelium (Lazarowski et al. 1992; Huang et al. 2001). The stimulated secretory response was pharmacologically identified to be the result of A_2B receptor activation and was lost in a cell line derived from a cystic fibrosis patient with a defect in ion transport at CFTR (CF/T43), implying that this ion channel is the one responsible for the A_2B-mediated Cl⁻ secretion (Lazarowski et al. 1992). It is worth noting that Calu-3 cells lack P2Y₂ receptors, the primary P2 receptor controlling ion transport in lung epithelium, which further implicates adenosine, and not ATP, as an important player in ion secretion. When this A_2B secretory response was compared to direct forskolin treatment, similar to what was found in the colonic epithelium, the A_2B pathway produced the same level of Cl⁻ secretion but a 9-fold lower level of cellular cAMP. This effect was attributed to tight colocalization of A_2B, adenylyl cyclase, PKA and CFTR (Huang et al. 2001). Further characterization of this mechanism revealed that the A_2B receptor is recruited to the plasma membrane upon stimulation and interacts with E3KARP (NHE3 kinase A regulatory protein) and ezrin (Sitaraman et al. 2002). Ezrin is a known PKA anchoring protein, and E3KARP associates with the COOH terminus of CFTR (Sun et al. 2000). These interactions may stabilize the receptor and downstream effector molecules in a signalling complex at the membrane. This compartmentalization of cAMP signalling challenges the idea of a readily diffusible second messenger and offers some understanding of how activation of pathways that utilize ubiquitous second messengers can function without global cellular activation.

While the A_2B-stimulated Cl⁻ secretory response was lost in the cystic fibrosis cell line CF/T43 specifically deficient in ion transport, manipulation of this receptor is still a consideration in treating cystic fibrosis patients with alternative CFTR mutations that preserve cAMP–PKA signalling and ion transport to some degree. For example, activation of the A_2B receptor can augment impaired Cl⁻ conductance at the R117H mutant CFTR, a channel which is known to localize to the cell surface and maintain normal PKA-dependent activation but which has reduced Cl⁻ conductance (Clancy et al. 1999). A_2B stimulation can also induce Cl⁻ secretion through the ΔF508 CFTR, where the defect is in trafficking to the membrane and in which Cl⁻ transport is preserved as long as the receptor spontaneously localizes to the cell surface a small percentage of the time or can be induced to localize by corrective molecules (Cobb et al. 2002). In Calu-3 cell monolayers, it was recently shown that phosphodiesterase inhibitors, specifically cilostazol, a phosphodiesterase 3 (PDE3) inhibitor, and papaverine, a non-specific PDE inhibitor, could induce a Cl⁻ current that was additive when combined with adenosine. This effect is presumably obtained by the elevation of intracellular cAMP. The concentration of cilostazol that increased chloride current was below therapeutic plasma levels achieved during the treatment of peripheral vascular disease (Cobb et al. 2003). While this has yet to be studied in cell lines with mutated CFTR channels, it raises the possibility that PDE inhibitor therapy alone or in conjunction with pharmacological manipulation of A_2B receptors could be of potential therapeutic benefit in cystic fibrosis.

A_2B receptors are not the only P1 receptors implicated in controlling ion transport across airway epithelium. Some literature also implicates the A₁ receptor, although the evidence for the role of this receptor has recently been called into question. RT-PCR studies have identified the A₁ receptor in the airway epithelial cell line A549, a normal tracheal cell line (9HTEo-), and a CF submucosal cell line (2CFSMEo-), distinguishing these epithelial lines from the Calu-3 cell line in which the A₁ receptor mRNA was not detected (McCoy et al. 1995; Szkotak et al. 2001, 2003). Antagonism at the A₁ receptor is thought to stimulate Cl⁻ efflux, as shown by the activity of the adenosine antagonist CPX in a variety of model systems (Schwiebert et al. 1992; McCoy et al. 1995; Haws et al. 1996). Importantly, the mechanism of action at this receptor seems to function even in the presence of a mutant CFTR gene, as shown by the ability of CPX to activate outward Cl⁻ currents in primary nasal epithelial cells from wild-type as well as homozygous ΔF508 CF patients (Schwiebert et al. 1992). This was confirmed by showing that CPX could induce iodide efflux from recombinant cells expressing ΔF508 CFTR (Haws et al. 1996). To explain this effect in cells that are otherwise impaired in Cl⁻ conductance, CPX must therefore work either to increase the Cl⁻ conductance of mutant CFTR channels that manage to reach the plasma membrane or augment impaired membrane trafficking in these cells to increase accumulation of these channels at the cell membrane. It has recently been shown that, despite the common assumption that CPX is an A₁ antagonist, this effect is probably due more to a direct interaction of CPX with CFTR than with the A₁ receptor. CPX has been shown to bind directly to the first nucleotide-binding fold domain (NBF-1) of wild-type CFTR and with even higher affinity to the NBF-1 of the ΔF508 CFTR, and recent evidence shows that CPX can activate wild-type recombinant CFTR chloride channels in an isolated lipid bilayer system (Cohen et al. 1997; Arispe et al. 1998). This evidence supports the idea that, despite the long-standing belief that A₁ antagonism activates cAMP-mediated Cl⁻ efflux via CFTR, the more important signalling event may actually be via a direct interaction of the antagonist with CFTR. The validity of CPX for use in treating CF patients has been preliminarily explored, specifically in two phase I clinical trials that tested the safety of CPX administration, although the efficacy of this treatment has yet to be determined (Noone et al. 2001; McCarty et al. 2002).

Although the A₁ receptor may not play a major role in regulating the cAMP-dependent CFTR chloride channel, it has been shown to regulate Ca²⁺-dependent processes in airway epithelial models. Specifically, in the wild-type human tracheal epithelial cell line 9HTEo-, the wild-type fetal trachea cell line 56FHTEo-, and the cystic fibrosis airway epithelial lines CFNPE9o- and CFPEo-, adenosine agonists have been shown to activate Ca²⁺-dependent K⁺ and Cl⁻ currents, similar to the action of ATP at P2Y₂ receptors in airway epithelia (Galietta et al. 1992; Rugolo et al. 1993). More recent work with the A549 airway epithelial cell line has confirmed the effect of A₁ activation on Ca²⁺-activated K⁺ channels, although the effect on Cl⁻ conductance has been questioned (McCoy et al. 1995; Szkotak et al. 2001).

P1 receptor signalling in renal epithelia

With regard to epithelial transport in the kidney, it has been known that adenosine stimulates sodium transport since Lang et al. (1985) first described it in Xenopus laevis renal epithelial A6 cells, a common model of the mammalian collecting duct. The net effect of adenosine stimulation is sodium transport from the lumenal to the serosal surface, making antagonists of these receptors possible candidates for diuretic therapy. Since this initial discovery, there has been much discussion about the specific adenosine receptor subtype involved in this response, with the A₁, A_2A and A₃ receptors all being implicated. There is evidence that A₁ stimulation, functioning through an increase in intracellular calcium and PKC activation, is largely responsible for apical stimulation of Na⁺ transport across the epithelial border (Macala & Hayslett, 2002). This is supported by the observed diuresis resulting from a decrease in sodium reabsorption thought to primarily occur in the proximal tubule following A₁ receptor blockade (Wilcox et al. 1999). In direct opposition to this is the observation that intrarenal delivery of A₁ agonists have been shown to induce natriuresis (Yagil, 1994). In addition, A₁ activation has been shown to inhibit the activity of the NHE3 Na⁺–H⁺ exchanger when transfected into A6 cells (Di Sole et al. 1999), and this effect could decrease net Na⁺ reabsorption along the renal tubule. These conflicting results, implicating A₁ agonism and antagonism in diuresis, highlight the ambiguity associated with P1 activation at the kidney epithelium. In addition to its effect on Na⁺ transport, stimulation of Cl⁻ secretion following A₁ receptor activation has been observed, specifically in the A6C1 subclone of the A6 cell line, and this secretory stimulation may account for some of the discrepancies in the observed effects of A₁ activation and antagonism (Casavola et al. 1996; Banderali et al. 1999; Macala & Hayslett, 2002).

The effects of A_2A receptor activation at the renal epithelium are equally debated. Basolateral A_2A activation is thought to activate transepithelial Na⁺ transport via a cAMP–PKA-dependent pathway, although the role of this receptor has been debated. It has been proposed that this regulation of Na⁺ transport occurs through A_2A-mediated control of intracellular pH via activation of the basolateral Na⁺–H⁺ exchanger and subsequent increase in intracellular pH (Casavola et al. 1997). The relationship between intracellular pH and Na⁺ transport is reciprocal such that the decrease in H⁺ concentration causes an increase in transepithelial Na⁺ transport. In contrast, A_2A agonists have been shown to inhibit the Na⁺–H⁺ exchanger NHE3 in transfected A6 cells (Di Sole et al. 1999, 2002).

Finally, it has also been proposed that A₃ receptors act at the apical surface to induce Cl⁻ secretion. This effect is dependent on a rise in intracellular Ca²⁺ but is thought to be via a novel G_s–PKA stimulated increase in calcium entry rather than the traditional G_i–PLC-β-mediated increase in intracellular calcium (Reshkin et al. 2000).

Adenosine plays a role in controlling renal function that is distinct from the direct effect on ion transport in epithelial cells. It has recently been shown that adenosine is a key regulator of tubuloglomerular feedback via activation of A₁ receptors on afferent arterioles (Schnermann & Levine, 2003). Accordingly, an A₁ receptor knockout mouse shows a loss of macula densa control of renal vascular tone (Sun et al. 2001). Taken together with the significant effects of adenosine receptors on renal Na⁺ and Cl⁻ transport, these data emphasize the importance of adenosine receptors in controlling the overall function of the kidney in both normal and pathogenic states, although further research is needed to resolve the conflicting reports of P1 function in the kidney.

P1 receptor signalling in other epithelial tissues

As previously discussed, the ubiquitous nature of purine nucleotide release by all cells in response to a variety of stimuli and the almost universal expression of purinergic receptors allows for P1- and P2-mediated control of ion transport in almost all epithelial subtypes explored. In accordance with this, P1 receptor control of ion secretion has been identified in such diverse tissue types as the non-pigmented ciliary epithelial cells of the retina, the vas deferens epithelium, and the middle ear epithelium.

The non-pigmented ciliary epithelial cells of the retina are a known location for A₃ receptor-mediated control of epithelial transport. A₃ activation causes an increase in aqueous humor production by increasing Cl⁻ secretion, leading to an increase in intraocular pressure. This has been supported by the observed increase in intraocular pressure following in vivo administration of A₃ receptor agonists as well as in the A₃ receptor knockout mouse model in which the intraocular pressure was reduced (Mitchell et al. 1999; Carre et al. 2000; Avila et al. 2001, 2002). Adenosine stimulates Cl⁻ channels of the ciliary epithelium, and this effect can be mimicked by administration of A₃ agonists such as N(6)-(3-iodobenzyl)-5′-N-methylcarbamoyladenosine (IB-MECA) (Carre et al. 1997, 2000; Mitchell et al. 1999). Transcript of the A₃ receptor has also been positively identified in human and rabbit ciliary epithelium by RT-PCR (Mitchell et al. 1999). It is also known that Cl⁻ efflux can be induced by cell swelling, and this current is dependent on the same Cl⁻ channels activated by A₃ receptor activation. It was previously thought that these shared a common downstream signalling pathway, possibly by the volume-sensitive release of ATP (Carre et al. 2000). The evidence for this was from rabbit non-pigmented ciliary epithelial cells in which a volume-sensitive phosphatidylinositol 3-kinase Cl⁻ efflux was dependent on activation of both PKC and PI3K, a common downstream pathway for G_i protein-coupled receptors such as the A₃ receptor (Selbie & Hill, 1998; Shi et al. 2002). Recently, this concept of a shared common pathway for volume-sensitive and A₃-induced Cl⁻ currents has been questioned due to the discovery that the A₃ receptor-induced Cl⁻ current in human ciliary epithelial cells is independent of PI3K but dependent on mitogen-activated protein kinase (MAPK) activation through a G_βγ pathway (Shi et al. 2003). This leads to the question of whether there are signalling differences between rabbit and human ciliary epithelium or whether adenosine induces Cl⁻ efflux signals through a separate pathway to that involved in volume-sensitive Cl⁻ efflux, despite being dependent on the same effector Cl⁻ channels (Shi et al. 2003).

With regard to the vas deferens epithelium, apical application of adenosine to primary human epithelial monolayers and freshly excised human vas deferens resulted in an increase in cAMP and subsequent anion (Cl⁻ and HCO₃⁻) release (Carlin et al. 2003). This recapitulates what has been shown in porcine vas deferens epithelia (Sedlacek et al. 2001). These results indicate that adenosine receptors modulate the lumenal environment of the deferent duct and may therefore play a role in male fertility. As in other tissue types, this has potential implications in treating cystic fibrosis, as over 97% of men with CF have reproductive dysfunction (Dean & Santis, 1994). In cultured gerbil middle ear epithelium, as in other tissues, adenosine has been shown to activate the cAMP–PKA system leading to Cl⁻ secretion through the CFTR Cl⁻ channels, and this effect is likely to be due to A_2B receptor stimulation (Furukawa et al. 1998). In addition to these tissues in which adenosine receptors have been directly implicated in the stimulation of anion currents, cAMP-dependent currents have been identified in many epithelial tissue types including epithelia from mandibular and submandibular glands, rectal glands, epididymis, oviduct, and endometrium, and further research is likely to indicate that activation of adenosine receptors, particularly the A_2B receptor, can stimulate a cAMP-mediated current in these tissues (Huang et al. 1993; Dinudom et al. 1995; Devor et al. 1995; Leung et al. 1995; Lee et al. 1999; Chan et al. 1999).

The role of P1 and P2 receptors in regulating ion transport at the epithelium is just one aspect of the diverse functions of these receptors. Further research needs to be initiated not only to further explore purinergic activity in epithelial tissues but also to determine the interactions between P1 and P2 receptors and between receptor functions in alternate tissue types. For example, the role of the A_2B receptor in inducing water secretion in response to an inflammatory stimulus in the colon may combine with the A_2A response on neutrophils recruited to the site to limit inflammation (Ohta & Sitkovsky, 2001). As more research on these receptors becomes available, it seems likely their manipulation will prove useful in clinical situations, especially in the treatment of cystic fibrosis and similar secretory defects.