HISTORIC ASPECTS OF PURINERGIC SIGNALING

TOP ABSTRACT HISTORIC ASPECTS OF PURINERGIC... PURINERGIC RECEPTORS:... ATP RELEASE INTO THE... SEGMENTAL DISTRIBUTION OF... LESSONS FROM THE STUDY... HYPOTHESES AND FUTURE... REFERENCES

SEVENTY YEARS AGO, Drury and Szent-Gyorgyi (41) were the first to recognize the potent extracellular actions of purine nucleotides and nucleosides in mammalian heart. After this initial report, research was limited to the actions of adenosine and ATP on the cardiovascular system (37, 70, 123). In the early 1960s, a component of the autonomic nervous system that was neither adrenergic nor cholinergic was identified in several tissues. Moreover, aspects of the cardiovascular system also had an unknown neurotransmitter that was neither norepinephrine nor acetylcholine (21, 24). In the early 1970s, Burnstock (21, 24) proposed that the principal neurotransmitter that is released from these specialized nerves was ATP. On the basis of the accepted criteria for neurotransmitters, Burnstock (21, 24) proposed the term "purinergic" for these nonadrenergic and noncholinergic nerves. After Burnstock presented his proposal, the concept of extracellular purinergic neurotransmission was strengthened by a larger body of experimental evidence demonstrating the role of ATP as a neurotransmitter or cotransmitter with norepinephrine, acetylcholine, or other chemical mediators. In addition, knowledge about specific extracellular receptors that mediate the physiological effects of purine nucleotides and nucleosides began to emerge.

In 1980, Burnstock proposed the classification of purinergic receptors into two major groups, P1 and P2, depending on the preferential affinity to adenosine or ATP, respectively (22, 23). This classification also took into account the selective activation of adenylate cyclase by adenosine and the induction of prostaglandin synthesis by ATP. Subsequently, Lodos et al. (86) and Van Calker et al. (147) identified two subclasses of adenosine (P1) receptors. These subclasses were later termed "A1" and "A2" in the nomenclature. By using molecular homology cloning methods (157) and receptor binding assays (33), A3 and A4 subtypes of P1 receptors were also identified. The A4 subtype is still not validated fully. These adenosine receptor subtypes couple to different heterotrimeric G proteins and effectors; adenosine receptors are not the major focus of this review (for reviews on adenosine receptors, see Refs. 100, 101, 108, and 133).

On the basis of the differences in rank-potency order profiles of nucleotides and nucleotide analogs, Burnstock and Kennedy suggested a subclassification of P2 receptors into P2X and P2Y subtypes (25). Later studies revealed that P2X and P2Y receptors also differ in their transduction mechanisms (55). P2X receptors appeared to have an intrinsic ion channel that increased the permeability of the plasma membrane to Na⁺, K⁺, and Ca²⁺ (and, possibly, anions), whereas P2Y receptors were traditional G protein-coupled receptors that coupled to heterotrimeric G proteins, phospholipases, and phosphoinositol signaling pathways. Gordon (50) further delineated the existence of P2T and P2Z subtypes of receptors into platelets and mast cells, respectively. These were originally thought to be distinctly different from the P2X and P2Y receptor subtypes. In his seminal review, Gordon also emphasized the concept of biologically relevant release of nucleotides and nucleosides, discussing the sources, effects, and fates of these purinergic agonists. Subsequently, another subtype of P2 receptor that responds to the pyrimidine nucleotide UTP as well as to ATP was identified and termed "P2U receptor" or the pyrimidine receptor. In addition to the P2 receptor subtypes discussed above, receptors that bind adenosine dinucleotide polyphosphates (Ap₄A, Ap₅A, and Ap₆A) were identified and classified as P2D receptors (27, 59). Although possible functions as neurotransmitters or cotransmitters have been suggested, the exact physiological roles of the diadenosine polyphosphates (Ap_xA) are not yet well established.

	PURINERGIC RECEPTORS: "METABOTROPIC" P2Y G PROTEIN-COUPLED RECEPTORS AND "IONOTROPIC" ATP-GATED P2X RECEPTOR CHANNELS

TOP ABSTRACT HISTORIC ASPECTS OF PURINERGIC... PURINERGIC RECEPTORS:... ATP RELEASE INTO THE... SEGMENTAL DISTRIBUTION OF... LESSONS FROM THE STUDY... HYPOTHESES AND FUTURE... REFERENCES

Figure 1 shows the probable topologies of P2Y and P2X purinergic receptors that we discuss below in this section.

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Purinergic receptor gene families and P2X receptor channel pharmacology and cation permeability. The physiological ligand for P2XRs and P2YRs is ATP itself as well as its metabolites. P2XR-specific agonists and less specific antagonists are shown. On gating by nucleotide ligand, P2X receptor channels are permeable to Ca2+ and Na+ entering the cell down large chemical gradients and to K+ leaving the cell down its chemical gradient. It is thought that ATP binds in the extracellular domain, although a binding site has not been mapped. Conserved cysteines in the extracellular domain likely form disulfide bridges and confer an elaborate secondary structure. In contrast, the structure of P2Y receptors is much like that of other G protein-coupled receptors. Meth, methylene; BzBz, benzoylbenzoyl; PPADS, pyrodoxal phosphate-6-azophenyl 2',4'-disulfonic acid; TM, transmembrane region; TNP, trinitrophenyl.

Metabotropic P2Y G Protein-Coupled Receptors

P2Y receptors vary in length from 308 to 379 amino acids, with a molecular mass of 41-53 kDa in glycosylated form. These receptors have seven transmembrane-spanning -helical hydrophobic regions typical of G protein-coupled receptors. The NH₂ terminus of the receptor proteins is on the extracellular side of the plasma membrane, whereas the COOH terminus lies on the cytosolic side of the membrane. The three-dimensional orientation of the transmembrane domains creates a pocket with positively charged amino acids that interact with the phosphate groups of the nucleotide ligands (ATP/ADP/UTP/diadenosine polyphosphates). These proteins are usually N-glycosylated on their second extracellular loop. Although the exact roles of glycosylation are not known, it has been suggested that the carbohydrate moieties stabilize the protein conformation, protect the receptors from the action of proteases, and modulate receptor function (for recent and exhaustive reviews, see Refs. 2 and 108).

The different subtypes of P2Y receptors interact with different types of G proteins through their intracellular loops. These G proteins usually activate a membrane-bound phosphatidylinositol-specific phospholipase C (PLC), resulting in enhanced formation of inositol 1,4,5-triphosphate (IP₃) and mobilization of cytosolic Ca²⁺ (Ca). In addition, activation of both P2Y₁ and P2Y₂ receptors leads to subsequent inhibition of adenylate cyclase through a G_i-dependent mechanism. Furthermore, the P2Y₁₁ receptor is unique among P2Y receptors because it couples to the stimulation of both phosphatidylinositol and adenylate cyclase signaling pathways. On the other hand, the activation of platelet P2Y_ADP receptor results predominantly in the inhibition of adenylate cyclase. Depending on the receptor subtype and on the cell type that expresses the receptor, the formation of IP₃ and mobilization of Ca can stimulate several other intracellular signaling pathways. These pathways include activation of protein kinase C (PKC), phospholipase A₂ (PLA₂), Ca²⁺-dependent K⁺ channels, or nitric oxide (NO) production. In addition, the diacylglycerol (DAG), formed as a result of PLC action, stimulates PKC, which, in turn, may activate phosphatidylcholine-specific PLC, phospholipase D (PLD), the MAP kinase pathway, and/or Ca²⁺ influx via voltage-operated Ca²⁺ channels. The specific downstream events seem to be dependent on the particular cell type in which the P2Y receptor is expressed. Thus in endothelial cells, the activation of P2Y₁ or P2Y₂ receptors results in the activation of PKC, without detectable elevation in IP₃ or Ca, and with subsequent rapid tyrosine phosphorylation of MAP kinase, resulting in the production of prostacyclins. Because of this complex nature of signal transduction mechanisms triggered by P2Y receptors, the response time of P2Y receptors is longer than the response time of P2X receptors (for recent and exhaustive reviews, see Refs. 2, 26, and 108).

P2Y receptors as a class do not desensitize as readily as other G protein-coupled receptor subfamilies. However, when they do desensitize, the mechanism involves either phosphorylation by protein kinases or uncoupling from the G protein. Suramin is a generalized antagonist of P2 receptors, except for P2Y₄. Presently, there are no specific antagonists available to distinguish the different subtypes of cloned mammalian P2 receptors, a void in the field that needs to be filled. Nonselective inhibitors such as suramin, reactive blue 2, and pyrodoxal phosphate-6-azophenyl 2',4'-disulfonic acid (PPADS) can be used to antagonize either or both of the P2Y and P2X receptors (16). PPADS blocks the P2Y₁ receptors coupled to PLC but not those coupled to the inhibition of adenylate cyclase. The platelet P2Y_ADP receptor is blocked by 2-propylthio-d---difluoromethylene ATP (FPL 66096) and 2-propylthio---dichloromethylene-d-ATP (ARL). P2Y receptor expression and function along the nephron will be addressed below (for recent and comprehensive reviews, see Refs. 2 and 108).

Ionotropic ATP-Gated P2X Receptor Channels

P2X receptors have no sequence homology to any other ion channel. In over 300 sequences, obtained by PCR amplifications on epithelial and endothelial cDNA libraries by using degenerate primers and subjected to the basic alignment research tool (BLAST) algorithm by Schwiebert and colleagues (Schwiebert LM, Rice WC, Kudlow BA, Taylor AL, and Schwiebert EM, unpublished observations; Schwiebert EM, Wallace D, King SR, Braunstein GM, Peti-Peterdi J, Hanaoka K, Guay-Woodford LM, Bell PD, Sullivan L, Grantham JJ, and Taylor AL, unpublished observations; 143), only P2X receptor channel subtypes are identified, with no other ion channel family emerging with even low or partial homology. Functionally, P2X receptor channels resemble glutamate receptors, another class of excitatory, ligand-gated receptor ion channels. Like glutamate receptors, P2X receptor channels are opened by a ligand and are permeable to monovalent and divalent cations. Divalent cations permeate as well as block the channel (16). Topologically, they are most similar to the amiloride-sensitive Na⁺ channels (ENaCs), inwardly rectifying K⁺ channels (IRKs, ROMKs), and nematode degenerins (16). Like ENaCs and IRKs, P2X receptors form heteromultimeric complexes. With the exception of P2X₆, they also form homomultimeric complexes (145). It is reasonable to assume that P2XRs form tetramers, given that the overall topology of the two transmembrane-spanning P2XR channels is similar to that solved by MacKinnon and colleagues (40) for a bacterial two-transmembrane-spanning K⁺ channel.

Table 2 summarizes the properties of the cloned P2X receptor channels. There are no specific agonists for the individual P2X receptor isoforms, but ,-methylene ATP, ,-methylene ATP, and benzoyl-benzoyl ATP will stimulate the P2X receptors without activating the P2Y receptors. Few specific antagonists exist that do not also inhibit P2Y receptors (see above), with the exceptions of trinitrophenyl ATP (TNP-ATP) and oxidized ATP (16). The seven P2X isoforms vary in their ATP-binding kinetics, conductances, and desensitization. The EC₅₀ values for ATP range from 0.7 to 15 µM for P2X₁ through P2X₆ (20). P2X₇ requires 300 µM of ATP to elicit a current (90). The pharmacology of P2X receptor channels has been reviewed recently in some detail (99). Single-channel conductances vary significantly from one subtype to the next. The permeabilities for sodium and potassium are similar, whereas Ca²⁺ permeability is much higher. Despite the higher permeability for divalent cations, Ca²⁺ and Mg²⁺ also block the channel. Of the seven isoforms, only P2X₁ and P2X₃ desensitize rapidly on binding to the agonist (20, 30). Interestingly, the most abundant isoforms expressed in epithelia (Schwiebert EM, Wallace D, King SR, Braunstein GM, Peti-Peterdi J, Hanaoka K, Guay-Woodford LM, Bell PD, Sullivan L, Grantham JJ, and Taylor AL, unpublished observations; 143) and endothelia (Schwiebert LM, Rice WC, Kudlow BA, Taylor AL, and Schwiebert EM, unpublished observations) are poorly desensitizing members of the P2X receptor family, P2X₄, P2X₅ and, to a lesser extent, P2X₂ and P2X₇. P2X receptor channel expression and function along the nephron will be addressed below.

	ATP RELEASE INTO THE EXTRACELLULAR MILIEU ALONG THE NEPHRON: ASSAYS, SOURCES, MECHANISMS, AND STIMULI

TOP ABSTRACT HISTORIC ASPECTS OF PURINERGIC... PURINERGIC RECEPTORS:... ATP RELEASE INTO THE... SEGMENTAL DISTRIBUTION OF... LESSONS FROM THE STUDY... HYPOTHESES AND FUTURE... REFERENCES

Extracellular Nucleotides as Agonists

View larger version (22K): [in this window] [in a new window]   Fig. 2.   ATP release mechanisms, extracellular ATP signaling, and emerging role of P2X receptor channels in the transduction of autocrine and paracrine ATP agonist signaling in airway or kidney epithelium. Release mechanisms and the large concentration gradient for ATP release from airway epithelia are shown on the left. P2X receptor channels in the apical or basolateral membrane of the same cell or a neighboring cell, on binding their ATP ligand, may increase cytosolic Ca2+ transiently and trigger Ca2+-dependent protein kinase signaling in a more sustained manner. These signaling cascades stimulate Cl and fluid secretion and ciliary beat frequency (shown in the model) as well as inhibit Na+ absorption and arginine vasopressin (AVP)-induced water reabsorption (not shown in the model).

ATP Release Mechanisms

Nonexocytotic release of ATP and other nucleotides likely occurs by passive transport mechanisms, using the large gradient for ATP efflux, secretion, or exit. Nonconductive transport of ATP (passively down its large concentration gradient out of the cell) may also be a route of ATP release. Such adenine nucleotide transporters have been studied intensely in mitochondrial membrane (19, 87, 119) and have been idenfied recently in endoplasmic and sarcoplasmic reticulum (53), chromaffin granule ghosts (7), and in rat brain synaptic vesicles (51). Nonconductive, bidirectional transporters exist in the plasma membrane for nucleosides primarily in brain (35); therefore, it cannot be discounted that ATP-specific transporters may exist. Indeed, the cystic fibrosis transmembrane conductance regulator (CFTR), once thought to be an ATP channel, may transport ATP as well as other large organic anions (gluconate, glutathione) at nonconductive rates. By the macropatch-recording method, Linsdell and Hanrahan (85) showed rates of transport of larger organic anions that were on the border between conductive and nonconductive.

ATP-permeable anion channels have been characterized biophysically by many laboratories. Their relationship to CFTR anion channels has been controversial (68, 83, 109, 110, 111, 124, 126); however, recent studies by Engelhart and colleagues (68), Foskett and co-workers (105, 136), and Schwiebert and colleagues (17) suggest that CFTR does not conduct ATP itself but regulates a closely associated anion channel that does conduct ATP. Anion channels with less selectivity for Cl vs. other halides or larger anions are prime candidates for putative ATP channels. These include, but are not limited to, the outwardly rectifying Cl channel (or ORCC) as well as plasma membrane forms of the voltage-dependent anion channel (VDAC; also called "porin") (10, 11, 112, 144). Large-conductance or "maxi" Cl channels that resemble mitochondrial VDAC or porin biophysically have been observed in most if not all cells. Interestingly, the major role of VDAC or porin in the mitochondrial membrane is the transport of newly synthesized ATP from the mitochondrion to the cytoplasm. Schwiebert and co-workers (142) as well as Fitz and colleagues (113, 114, 115, 149) have shown that ATP release is stimulated under conditions of hypotonic stress. This release is immediate (within seconds), and it precedes the time course for regulatory volume decrease after cell swelling (17, 142). Another elegant example of a paracrine ATP signal involves propagation of Ca²⁺ waves along monolayers of cells in a gap junction-independent manner (49, 120). Thus these are key examples of how ATP works well as a signaling molecule, because it is released rapidly down its large concentration gradient, it acts in an autocrine or paracrine manner, it is rapidly degraded to dissipate the response, and it mediates fast and slow responses, via P2X and P2Y receptors, respectively (16).

Assays to Detect ATP Release and Signaling

Another elegant assay determines ATP release by using luciferin-luciferase attached to the surface of a cell. Luciferase is fused to protein A, a protein that binds to IgG antibodies. Cells are incubated first with an IgG antibody to an extracellular epitope of a specific cell surface protein, then with the protein A-luciferase fusion product. The cells are placed in a luminometer to collect photons as an indicator of localized extracellular ATP release at the membrane surface (9). This assay was utilized on hematopoietic cells, where surface antigens are abundant and well characterized. However, it could be adapted to tissue preparation such as isolated and perfused renal tubules, intestinal crypts, or isolated glands or ducts.

The most recent assay involves atomic force microscopy (AFM) using commercially available AFM tips coated with the myosin subfragment S1, which has a high affinity for ATP and changes shape on ATP hydrolysis. Rather than luciferase-luciferin reagent as the readout, the myosin tips are placed next to a cell, bind the ATP as it is released, and the myosin in the tips responds by changing shape (121). Using this assay, Schneider and co-workers (121) showed that cystic fibrosis airway epithelial cells released little ATP (e.g., the vibration of the AFM probe was minimal); however, when the probe was placed in close proximity to the membranes of cystic fibrosis (CF) cells stably complemented with wild-type CFTR, significant vibration of the probe was observed. Adaptation to isolated tubule preparations has not been performed; however, it is a definite possibility.

Creative indirect assays have also been developed, using the P2X receptor channels as a functional readout. Hazama et al. (57) developed an assay that determines ATP release from a single cell. A single PC12 cell, which expresses P2X₂, is lifted while in whole-cell patch configuration and placed near a pancreatic -cell growing on a coverslip. The pancreatic -cell is then stimulated with glucose, while the PC12 cell is studied for P2X₂ current. Current is indicative of the -cell's release of ATP (57). This assay is applicable to any cell or dissected tissue preparation of interest (see below), when maintained in culture. Moreover, instead of P2X receptor current as an endpoint, one could load the PC-12 cell with fura 2-acetoxymethyl ester (AM) and measure Ca²⁺ influx through P2X₂ as a different fluorescent endpoint (see below). Hollins and Ikeda (60) used a similar strategy. They transfected rat adrenal chromaffin cells with P2X₁ and then stimulated the cells with agents that increase exocytosis. Their patch-clamp data showed that stimulating exocytosis led to P2X₁ current, implying that ATP is released in a vesicular manner. Indeed, other laboratories are using these groundbreaking assays or modifying them further with great success. Moreover, their possible application to the kidney and to cell or tissue preparations of the nephron was reviewed recently in more detail (125) and will be expanded on below (see HYPOTHESES AND FUTURE DIRECTIONS).

Putative ATP Release and Release Mechanisms Along the Nephron

Figure 3 shows estimated concentrations of ATP based on parallel standard curves using the same luciferase-luciferin-containing reagent but with known quantities of ATP. The proximal tubule is the richest source of ATP under basal conditions, at concentrations that reach ~1 µM. On stimulation, this concentration increases; however, maximal ATP release never exceeds 5-10 µM. A study using human renal primary cultures to document ATP concentrations measured under basal and hypotonic conditions with this bioluminescence assay have been published (151). As other cultures are examined along the nephron, the amount of ATP release under basal conditions decreases through the nanomolar range. Collecting duct epithelia release low-nanomolar to picomolar concentrations of ATP under basal conditions, but collecting duct cell lines can be induced to release ATP with hypotonicity (a condition that would be present in diuresis luminally) and with the Ca²⁺ agonists ionomycin and thapsigargin. (Schwiebert EM, Wallace D, King SR, Braunstein GM, Peti-Peterdi J, Hanaoka K, Guay-Woodford LM, Bell PD, Sullivan L, Grantham JJ, and Taylor AL, unpublished observations). As a general principle, ATP release into the apical medium is more robust than basolateral-directed ATP release (151). Hypotonic challenge stimulates ATP release immediately and with transient and sustained components of a time course across the apical and basolateral membranes. Ca²⁺ agonists stimulate a slow, monophasic rise in ATP release that plateaus after several minutes. Ca²⁺ agonists stimulate apical-directed ATP release, whereas they are without effect on basolateral release. Present work is focused on studying the regulation of ATP release as well as ATP release mechanisms in renal epithelial cell models. Similar work is ongoing in human vascular endothelial monolayers (Schwiebert LM, Rice WC, Kudlow BA, Taylor AL, and Schwiebert EM, unpublished observations). The ultimate goal is to adapt this assay to the study of whole kidney or isolated nephron segments (see below).

View larger version (27K): [in this window] [in a new window]   Fig. 3.   Amounts of ATP released by different nephron segments under basal and stimulated conditions: "ballpark" concentrations. The results, accumulated by Schwiebert and colleagues, are based on bulk or total ATP release measurements from cell lines and primary cultures derived from known nephron segments and grown in vitro in 35-mm dishes. The amount of ATP release was correlated with a standard curve with known concentrations of MgATP. Amounts shown are estimates of what might be achieved in vivo under stimulated conditions; however, all cell types studied released significant amounts under basal conditions as well. CCD, cortical collecting duct; TLs, thin limbs; mTAL, thick ascending limb of Henle's loop; OMCD and IMCD, outer medullary and inner medullary collecting duct, respectively; PCT and DCT, proximal and distal convoluted tubule, respectively; PST, proximal straight tubule.

Bell and colleagues have used the PC-12 cell biosensor method developed by Okada and co-workers to detect ATP release across the basolateral membrane of the macula densa (MD) plaque that lies within the cortical thick ascending limb (cTAL) juxtaposed to the afferent arterioles and the glomerulus (Bell PD, personal communication). It is their hypothesis that ATP is released from macula densa (MD) in response to cTAL transport of Cl across the MD as an autocrine/paracrine signal for tubuloglomerular (TG) feedback. With the use of whole cell patch-clamp recording of PC-12 cells placed near the basolateral surface of the MD plaque, ATP-gated P2XR current was detected, indicative of ATP release from the plaque. Similarly, PC-12 cells loaded with the Ca²⁺-sensitive dye fura 2-AM were used as a fluorescence biosensor, and an increase in Ca was detected, also indicative of physiologically relevant and sufficient ATP release to stimulate receptors on the biosensor cell in vitro or in the glomerulus in vivo. In both cases, ATP release was stimulated by low NaCl in the lumen of the cTAL and was blocked by luminal and basolateral Cl transport inhibitors. To address the mechanisms of ATP release, cells of the MD plaque were patch-clamped and the ATP permeability of anion channels recorded in this membrane was examined. The most promising candidate for an ATP-permeable anion channel was a maxianion channel of ~300 pS that had an ATP conductance of ~100 pS and also had permeability to other anions such as Cl and gluconate (Bell PD, personal communication). Intriguingly, the voltage dependence of this ATP-permeable channel is similar to the VDAC. As described above, the role of VDAC or porin in mitochondrial membrane is to transport newly synthesized ATP out of the mitochondrion into the cytoplasm (10, 11). If plasma membrane-expressed VDAC channels exist (112, 144), this may be a principle ATP release channel. Nevertheless, this is intriguing preliminary work and is an example of a physiological role for extracellular ATP release and signaling.

Kishore and Knepper incubated freshly microdissected rat inner medullary collecting duct (IMCD) segments in vitro under physiological conditions at 37°C in Terasaki plates under mineral oil and with oxygenation. They measured the release of ATP by the tubules into the medium at different time points using the luciferin-luciferase assay system in a luminometer. Their data showed that that IMCD segments release ATP at a rate of 2.11 ± 0.29 fmol · mm¹ · 30 min¹, with a range of 1.43-3.3 fmol. This release was confirmed further in microperfused IMCD, which released ATP into the lumen at a rate of 1.52 fmol · mm¹ · 10 min¹. They also correlated this result with the total ATP content of the IMCD segments. The ATP content was 1 log order of magnitude greater than the rate at which the IMCD released ATP. They concluded that the release was physiological and did not represent a nonspecific leakage due to cell death. These preliminary data indicate that ATP can be released by the collecting duct cells as an autocrine, paracrine, or autoregulatory mediator (Kishore BK and Knepper MA, unpublished observations).

More indirectly, two laboratories have documented a role for endogenous ATP release and signaling as a modulator of intracellular signaling set points. Insel and co-workers (63) have recently published the conclusion that ATP, released endogenously and continuously by Madin-Darby canine kidney (MDCK) cells, modulates phosphatidylinositol signaling and turnover as well as cAMP production in MDCK cells. They showed that removal of the endogenously released ATP with the ATPase/ADPase apyrase or antagonism of P2 receptors significantly decreased arachidonic acid release as well as cAMP production. They also observed that vigorous medium changes on MDCK and other cultures stimulated ATP release mechanically. Similar results have been observed in fura 2-AM Ca imaging experiments by Schwiebert and colleagues (Schwiebert EM, Wallace D, King SR, Braunstein GM, Peti-Peterdi J, Hanaoka K, Guay-Woodford LM, Bell PD, Sullivan L, Grantham JJ, and Taylor AL, unpublished observations). In a circulating system where basal Ca was monitored in normal and polycystic kidney [polycystic kidney disease (PKD)] primary cultures, the ATP scavengers hexokinase and apyrase or the P2Y and P2X receptor antagonist suramin lowered basal Ca²⁺ significantly and reversibly. In fact, 10 and 100 µM suramin dropped basal Ca significantly in a rapid and reversible manner (Schwiebert EM, Wallace D, King SR, Braunstein GM, Peti-Peterdi J, Hanaoka K, Guay-Woodford LM, Bell PD, Sullivan L, Grantham JJ, and Taylor AL, unpublished observations). Taken together, these two studies suggest that autocrine and paracrine ATP signaling occurs constantly in the extracellular milieu and, at least in part, establishes a "set point" for multiple signal transduction pathways or signaling molecules. Indeed, basal modulation of cell volume regulation and Cl channel activity by endogenous ATP signaling was also shown in hepatocytes and cholangiocytes by Fitz and co-workers (114).

Taken together, for purinergic receptor expression to be relevant physiologically (see next section) there must be biologically relevant ATP release occurring in the same tissue, culture, or microenvironment. Not as much attention has been paid to ATP release in the past; however, interest in this aspect is gaining momentum in the purinergic receptor field. It is likely that these fields will merge and synergize to bring the larger picture of autocrine/paracrine extracellular purinergic signaling into focus.

	SEGMENTAL DISTRIBUTION OF PURINERGIC RECEPTORS ALONG THE NEPHRON EPITHELIUM: PURINERGIC SIGNALING AND REGULATION OF RENAL EPITHELIAL FUNCTION

TOP ABSTRACT HISTORIC ASPECTS OF PURINERGIC... PURINERGIC RECEPTORS:... ATP RELEASE INTO THE... SEGMENTAL DISTRIBUTION OF... LESSONS FROM THE STUDY... HYPOTHESES AND FUTURE... REFERENCES

The physiological significance of purinergic signaling in different cell types along the nephron and in specific nephron segments is an emerging and complex area of research. The principles of ATP release, ATP degradation, multiple subtypes and subfamilies of ATP receptors, and their coupling to multiple signal transduction cascades (discussed above) are complex in itself. This complexity is magnified by the 20 or more different cell types along the nephron and the 10 or more discrete nephron segments that are in contact with the tubular fluid or interstitium containing nucleotide or nucleoside agonists. In this section, themes will emerge that presently guide this field and will continue to guide it for some time to come. First, purinergic receptors are expressed abundantly in the kidney on all cell types of the glomerulus, on vascular smooth muscle and endothelial cells of the renal vasculature, and on the renal epithelium in all nephron segments where it has been examined. Second, in a given renal epithelial cell model (MDCK, mIMCD-K2, etc.), multiple P2Y receptors and multiple P2X receptor channels are expressed in the same epithelial cell and, often, in the same membrane domain of the epithelial cell. The reason this redundancy in ATP receptors is needed is an important fundamental question. Redundancy may be needed for an essential physiological process in all cells such as cell volume regulation, or the P2Y receptors and P2X receptor channels may fulfill completely different roles in the same epithelial cell. Third, and most important, how does purinergic signaling as an autocrine and paracrine agonist cascade regulate renal function independently of other agonists and how does it modulate the effects of important endocrine hormones such as vasopressin or angiotensin II or aldosterone? Each of these fundamental questions is just beginning to be addressed.

P2Y Receptors Along the Renal Epithelium

Kishore et al. (73) mapped P2Y₂ receptor expression along the mammalian nephron. Specific antibodies to the P2Y₂ receptor as well as a gene-specific cDNA probe confirmed its expression in terminal IMCD as well as its ubiquitous expression in all regions and nephron segments of the kidney. Apart from the medullary regions and segments, cortical regions and segments also showed strong immunohistochemical signal (Kishore BK, unpublished observations). Of interest, expression of the P2Y₂ receptor on both membrane domains of the IMCD was also found (see Fig. 4), which is similar to the expression of vasopressin V₂ receptor in IMCD (98). This is both interesting and perplexing, and this issue requires further investigation. The present working model for the regulation of collecting duct water permeability that grew out of the work of Kishore and co-workers as well as other investigators is shown in Fig. 4. AVP, acting through its V₂ receptor and the cAMP second messenger system, increases osmotic water permeability of the collecting duct apical membrane by translocating the aquaporin-2 (AQP2) water channel-containing vesicles from a subapical pool to the apical plasma membrane. The apical membrane is the rate-limiting barrier for transepithelial water transport, as AQP3 and AQP4 are expressed constitutively in the basolateral plasma membrane under normal conditions (for reviews, see Refs. 96 and 97). On the other hand, the agonist activation of the P2Y₂ receptor for extracellular nucleotides, as well as the endothelin receptor (ET_BR) and the EP₃ subtype receptor of prostaglandin E₂ (EP₃R), antagonize the AVP-stimulated water transport in collecting duct (18, 43, 58, 71, 79, 116). This antagonism is achieved by virtue of the cross-talk mechanisms that exist between the two mutually opposing intracellular signaling pathways in the collecting ducts, as depicted in Fig. 4 (18, 58, 71, 78, 116).

View larger version (34K): [in this window] [in a new window]   Fig. 4.   IMCD cell model showing purinergic and other local agonist receptor-mediated inhibition of AVP-induced apical membrane water permeability. Schematic representation of the two mutually opposing intracellular signaling pathways and the corresponding membrane receptors involved in the regulation of osmotic water permeability of the collecting duct principal cell. For details, see the text. The nature of the G proteins associated with the endothelin B receptor (ETB-R) in vivo in collecting duct is not established conclusively. Gi, inhibitory G protein; Gs, stimulatory G protein; PKC, protein kinase C; AC, adenylate cyclase; PLC, phospholipase C; cPLA2, cytosolic phospholipase A2; IP3, inositol trisphosphate; PKA, protein kinase A; DAG, diacylglycerol; PDE, phosphodiesterase; AQP2, aquaporin-2.

By measuring Ca activity in terminal IMCD after exposure to various nucleotide analogs, Ecelbarger et al. (42) demonstrated the existence of a classic P2Y₂ purinoceptor in the rat terminal IMCD (42). Suko et al. (137) reported similar results showing that extracellular ATP induces increases in both Ca and nuclear Ca²⁺ in cultured IMCD cells (137). McCoy et al. (92) showed functional and RT-PCR evidence for expression of P2Y₂ and P2Y₁ receptors in mIMCD-K2 cells. Cha and co-workers (29) also showed evidence for the existence of P2Y₂ as well as P2Y₁ receptors in rat outer medullary collecting duct (OMCD) segments (29). Kishore et al. (71) extended the work of Ecelbarger et al. in Knepper's group (42) to show that ATP and UTP applied to the basolateral surface of rat terminal IMCD inhibited AVP-induced osmotic water permeability (P_f) in microperfused IMCD. McCoy et al. (92) also showed functional effects of P2Y₂ receptor activation in IMCD. UTP, ATP, or methylene-ATP analogs that stimulate P2X receptor channels (see section below) stimulated Cl secretion across monolayers of mIMCD-K2 cells. Moreover, in the same monolayers, these nucleotide agonists inhibited Na⁺ absorption. Taken together, these studies show that P2Y₂ receptor localization and activation in medullary collecting duct segments have a profound effect on cell signaling and modulation of salt and water transport across these segments.

Recently, Deetjen and colleagues (36) have performed elegant fluorescence imaging studies in isolated perfused mouse cortical collecting duct (CCD) showing that luminal UTP triggers an increase in Ca. Comparison of the response of many nucleotide analogs led them to conclude that the P2Y₂ receptor mediated this response. In the course of their studies, they showed that serosal addition of nucleotide agonists also increased Ca; however, the rank order potency of the agonists was very different. The basolateral P2 receptor in mouse CCD may be a different purinergic receptor entirely. Satlin and co-workers (152) have made similar observations in rabbit CCD; moreover, they have found luminal effects of ATP scavengers and ATP receptor antagonists on ion transport in isolated perfused and impaled rabbit CCD (L. Satlin, personal communication). Rouse and colleagues (116) also showed antagonism of AVP-induced water transport in rabbit CCT. Examination of a large panel of agonists showed that ATP, ADP, UTP, and methylene-ATP analogs attenuated AVP effects, suggesting the expression and involvement of multiple P2Y and P2X receptors in their inhibitory effects (116). Lu and co-workers (88) have shown inhibition of small-conductance, secretory K⁺ channels in freshly isolated, "filleted open" CCD by purinergic agonists (88). In findings to those of Rouse and co-workers (116), ADP and UTP each inhibited the channels, suggesting that multiple P2Y receptors (P2Y₂ and, possibly, P2Y₁) may underlie this regulation.

Utilization of heterologous renal epithelial models derived from kidney, like the distal nephron models in MDCK cells, A6 Xenopus laevis kidney cells, and LLC-PK₁ porcine kidney cells, has provided ideal models in which to study purinergic signaling. As early as 1979, Simmons (129, 130) showed that exogenous ATP agonists affect ion transport across MDCK monolayers. Soon after this finding, Simmons and co-workers (131) showed that extracellular ATP stimulated Cl secretion across MDCK monolayers. Very recently, they revisited this work in mIMCD-K2 cells and found that external ATP stimulated Cl secretion in a similar manner (14). However, it is Insel and co-workers (48, 64, 108) who have led the way in this regard with multiple publications regarding the MDCK model. They have examined purinergic signaling via phosphoinositide signaling cascades. Arguably, their work has triggered present work in isolated nephron segments or in epithelial cell models derived from defined nephron segments. Rather than discuss their work in detail, Insel (63) has recently published a review on his work that does better justice to this body of literature. Zambon and co-workers (155) have shown recently that P2Y₁, P2Y₂, and P2Y₁₁ are expressed in MDCK cells, and this laboratory has cloned the canine forms of these P2Y receptor genes (P. A. Insel, personal communication). In addition to work by Insel and co-workers, Kishore, Ecelbarger, and co-workers have done similar work on nucleotide agonist-dependent signaling in IMCD. Ecelbarger et al. (42) demonstrated that prior exposure of rat IMCD to indomethacin, an inhibitor of cyclooxygenase (COX), attenuates Ca response to ATP agonists. This observation suggested that an as yet unidentified COX by-product of arachidonic acid metabolism (e.g., prostaglandin E₂) is facilitating or mediating the Ca response to ATP. Indeed, preliminary studies by Kishore and co-workers (75) revealed that ATP or UTP, but not ADP, enhanced the production and release of PGE₂ in a COX- and MAP kinase-dependent manner. These studies underscore the theme in purinergic signaling in the kidney and elsewhere that purinergic-triggered signaling pathways are quite complex. This complexity is further enhanced by the presence of multiple P2Y (and P2X) receptors in the same epithelial cell and, in some cases, in the same membrane domain of an epithelial cell.

Indeed, evidence for expression of additional P2Y receptor subtypes is accumulating. In a few studies highlighted above, P2Y₁ expression along with P2Y₂ was documented. In the S1 segment of PCT as well as in OMCD, stimulation by a panel of nucleotide agonists suggested the concomitant expression of P2Y₁ and P2Y₂ (29). In a parallel study by the same laboratory, RT-PCR in cell lines derived from S1 PCT and OMCD showed expression of P2Y₁, P2Y₂, and the P2X receptor channel subtype P2X₄ (140). Rank order potency studies of nucleotide agonist effects on Ca in rabbit renal proximal tubule showed that P2Y₁, but not P2Y₂, was expressed in this segment and mediated ATP-induced increases in Ca (154). Jin and Hopfer (69) found similar results in a renal cell line derived from early proximal tubule of the rat. Specific RT-PCR and functional studies also showed expression of both P2Y₁ and P2Y₂ in mIMCD-K2 cells (92). Rouse and co-workers (116) had pharmacological evidence that P2Y₁, P2Y₂, and P2X receptor channels may all mediate ATP-mediated attenuation of AVP-induced water permeability in rabbit renal CCT (116). Bouyer et al. (15) also found a P2Y₁-like phenotype in Necturus proximal tubule that increases Ca and stimulated a basolateral Cl channel. Finally, Ishikawa et al. (65) found that both P2Y₁ and P2Y₂ receptors may mediate increased growth rates of rat renal IMCD cells in vitro. Taken together, P2Y₁ alone or together with P2Y₂ may mediate effects on ATP along the nephron in multiple nephron segments, specifically proximal tubule and collecting duct.

Very recently, Bailey and co-workers (4) have added a third P2Y receptor to the list of those subtypes expressed along the rat renal epithelium. Single-nephron-segment RT-PCR revealed that expression of P2Y₆ is high in PCT and in the thick ascending limb and thin descending limb of Henle's loop. Expression was qualitatively lower in OMCD, although RT-PCR is not precisely quantitative. Expression was absent in the thin ascending limb of Henle as well as in IMCD. Functional expression was assessed by using basolateral perfusion of the selective nucleotide UDP, which stimulates P2Y₆ receptors more readily than it does other subtypes. UDP-triggered increases in Ca were robust in PCT, weak in OMCD, and absent in other segments. UDP-stimulated PI turnover was also examined and was present in PCT and OMCD but not in other segments. These results are the first to describe molecular and functional evidence for expression of the P2Y₆ receptor along the nephron. Taken together, these studies show that multiple P2Y receptors are poised along the nephron to transduce the autocrine and paracrine ATP signal as it travels down the nephron.

P2X Receptor Channels Along the Renal Epithelium

Recently, Schwiebert and co-workers (Schwiebert EM, Wallace D, King SR, Braunstein GM, Peti-Peterdi J, Hanaoka K, Guay-Woodford LM, Bell PD, Sullivan L, Grantham JJ, and Taylor AL, unpublished observations) have attempted to map P2X receptor channel expression in cell line and primary cultures derived from defined nephron segments. Degenerate RT-PCR with primers that amplify all seven isoforms of the P2X receptor gene family was performed. Verification of this work is ongoing with subtype-specific antibodies. The amplified product was subcloned for sequencing, and 12-32 white bacterial colonies bearing a PCR product insert were sequenced for each epithelial cell mRNA sample. Each sequence derived from each white bacterial colony was subjected to the BLAST algorithm to define and confirm the identify of the P2X receptor subtype found (3). In human mixed renal epithelial primary cultures, two different proximal tubule primary cultures, in human mesangial cell primary cultures, and in human autosomal dominant polycystic kidney disease (ADPKD) primary cultures, P2X₄ and P2X₅ are expressed abundantly to the exclusion of all other isoforms. Only human renal mixed epithelial and human proximal tubule primary cultures expressed a third isoform, P2X₇, and only with low incidence. In the RCCT-28A cell line, a model of A-type intercalated cells, P2X₅ was the predominant subtype expressed, whereas, in mIMCD-K2 cells, P2X₃ and P2X₄ were expressed at equal incidence to the exclusion of other isoforms. In collecting duct primary cultures from wild-type mice and the cpk autosomal recessive PKD mouse, P2X₃ was the most abundant sequence found, whereas a smattering of P2X₁, P2X₂, and P2X₄ was also found in much lower incidence. These recent and novel findings are shown in a schematic fashion in Fig. 5. They need to be confirmed by biochemical methods as well as single-nephron RT-PCR with degenerate primers and with primers specific to each subtype; however, they agree with studies in epithelial cell models from other tissues as well as the early studies described above. Taken together, these results suggest that multiple P2X receptor channels, in addition to multiple P2Y receptors, are poised to receive autocrine/paracrine nucleotide agonist signals as they travel along the renal epithelium.

View larger version (23K): [in this window] [in a new window]   Fig. 5.   P2X receptor channel expression along the renal epithelium. The figure is based on degenerate and/or specific RT-PCR performed on cell lines derived from defined nephron segments or mesangial cells isolated from glomeruli. Because subtype-specific antibodies are only recently available, this work needs to be confirmed by biochemistry, immunocytochemistry, and single-nephron PCR.

Pathophysiological Paradigms in Which Purinergic Signaling May Be Beneficial or Detrimental

Hypertension. One major area in which purinergic signaling may have import is that of hypertension. Because extracellular purinergic signaling modulates NaCl and water handling along the nephron, this signaling could affect certain hypertensive states. Moreover, because salt and water reabsorption is inhibited by extracellular ATP in the tubular lumen or apical environment, delivery of ATP as a therapeutic agent could lower salt and water reabsorption, blood volume, and blood pressure. Certainly, studies by Kishore et al. (71) as well as McCoy et al. (92) agree and underscore this hypothesis. On the other hand, Churchill and Ellis (32) demonstrated that, in rat renal cortical slice preparation, ATP (100-500 µM) and its analogs stimulate renin secretion in a concentration-dependent manner. They also demonstrated that this effect is mediated by a subtype of P2Y receptor via nitric acid (32). Thus it is likely that the multiple purinergic mechanisms may play a role in blood pressure regulation through kidney.

Water balance. Kishore and co-workers (72) have extended their work on the involvement of purinergic signaling in rat models of hypo- and hypervolemia and polyuria of ischemic-reperfusion injury. Preliminary data indicate that the abundance of the P2Y₂ purinoceptor protein in the renal medulla is increased significantly in the hydrated state (hypervolemia) compared with dehydrated state (hypovolemia). Furthermore, there is an apparent shift in the subcellular localization of this protein in the medullary collecting duct cells, as revealed by immunoperoxidase labeling with P2Y₂ receptor-specific antibody. In the hypovolemic condition, a predominantly basal labeling is seen as opposed to a predominantly apical labeling in the hypervolemic state (72). Interestingly, Kishore and associates (74) also observed that the distribution of immunoreactive cytosolic phospholipase A₂ (cPLA₂) in the medullary collecting duct cells is also altered under these conditions, with an intense labeling for cPLA₂ on the basal aspect of the cells. These preliminary observations, when established further by more extensive studies, will underscore the importance of the association between the purinergic signaling and arachidonic acid metabolism in the long-term conditioning of collecting duct water permeability and thus offer a basis for the vasopressin-independent regulatory mechanisms. Furthermore, in a preliminary study Kishore and colleagues (76) observed that, in the renal medulla of rats during the reperfusion phase after bilateral renal pedicle clamping, P2Y₂ purinoceptor mRNA was markedly increased with a concomitant decrease in AQP2 water channel mRNA. Because it has been well documented by Fernandez-Llama and co-workers (45) that decreased aquaporin protein abundance in collecting duct cells is a contributing factor in the increased urine flow seen in moderate postischemic acute renal failure, the preliminary observations of Kishore et al. (76) on P2Y₂ expression suggest the probable involvement of this receptor in the diuretic condition of ischemic reperfusion injury (IRI). Moreover, release of ATP and adenosine in other tissues during IRI has been shown and could play additional modulatory roles beyond regulation of water balance by the medulla.

Tubuloglomerular feedback. Studies by Bell and colleagues have shown that ATP signaling in the interstitium between the macula densa plaque in the cTAL and the afferent arteriole-glomerulus complex may aid in tubuloglomerular feedback mechanisms that regulate renal blood flow and glomerular filtration rate (Bell PD, personal communication). In pathophysiological states where these parameters change dramatically, autocrine/paracrine ATP signaling may be affected.

PKD. Alternatively, in recessive and dominant forms of PKD, extracellular ATP signaling could prove detrimental. Wilson et al. (151) showed that human PKD epithelial monolayers release as much or more ATP than do normal epithelial controls. They also showed that a subset of human PKD cyst fluid samples had nanomolar to micromolar ATP concentrations (some as high as 10 µM ATP). All samples had measurable ATP levels above background, suggesting that ATP release into the PKD cyst lumen and extracellular ATP signaling within the cyst occur in vivo. They have expanded this data set into other PKD models more recently (Schwiebert EM, Wallace D, King SR, Braunstein GM, Peti-Peterdi J, Hanaoka K, Guay-Woodford LM, Bell PD, Sullivan L, Grantham JJ, and Taylor AL, unpublished observations). Figure 6 shows our general working hypothesis concerning the possible detrimental impact of extracellular purinergic signaling. ADPKD cysts, as fluid-filled structures or spheres lined by a single layer of epithelial cells, present a physiological enigma. Ion and fluid transport become encapsulated, especially secretory transport into the cyst. As such, any secretagogue may be detrimental. Indeed, growth factors are released and trapped with the cyst lumen where they interact with growth factor receptors, creating a vicious autocrine/paracrine growth loop (95, 150). Purinergic agonists are mitogens or comitogens with growth factors in some cell models (44, 49, 103, 106, 120, 123, 148).

View larger version (23K): [in this window] [in a new window]   Fig. 6.   ATP release, purinergic receptors, and purinergic receptor-mediated signaling may be detrimental in polycystic kidney disease (PKD). This schematic shows how an autosomal dominant (AD) PKD cyst arises slowly from a normal renal tubule to become a fluid-filled sphere encapsulated by a single cell layer of renal epithelial cells. ADPKD cysts arise heterogeneously from the nephron but mostly from distal nephron. Like growth factors, ATP and its metabolites are "trapped" as nucleotide and nucleoside agonists in the cyst fluid. Trapped ATP and ATP metabolites, as fluid and Cl secretagogs and mitogens or co-mitogens with already trapped growth factors, may exacerbate fluid accumulation and volume expansion in ADPKD cysts and proliferation of epithelial cells that line ADPKD cysts. The hypothesized result is slowly accelerated cyst expansion and growth. Purinergic receptors of all types are expressed on renal epithelium; thus the autocrine signaling cascade may be intact in the lumen of the cyst. Finally, recent work from Schwiebert and co-workers has shown that nucleotide agonists increase cytosolic Ca2+ in ADPKD cells, whereas collaborative work with Wallace, Sullivan, and Grantham shows that nucleotide agonists stimulate Cl secretion across human ADPKD monolayers grown in primary culture. Clearly, all of the elements exist for purinergic stimulation of Cl secretion via cytosolic Ca2+ signaling, a function that is detrimental when the nephron architecture has changed from a tubule to a cyst. Ado, adenosine.

	LESSONS FROM THE STUDY OF EPITHELIA DERIVED FROM OTHER TISSUES

TOP ABSTRACT HISTORIC ASPECTS OF PURINERGIC... PURINERGIC RECEPTORS:... ATP RELEASE INTO THE... SEGMENTAL DISTRIBUTION OF... LESSONS FROM THE STUDY... HYPOTHESES AND FUTURE... REFERENCES

A multitude of laboratories have studied purinergic agonists and their effects of epithelial cell models. In particular, purinergic agonists were among a large panel of agonists tested for their ability to stimulate Cl and fluid secretion from cystic fibrosis (CF) tissues and epithelial cell models from the lung and airways and from the gastrointestinal (GI) system. In CF, Cl and fluid secretion are lacking, whereas sodium absorption is augmented. Purinergic agonists were among the few agonists tested that were successful in correcting this defective NaCl transport in CF cells (77, 80, 134, 135). As such, UTP and UTP analogs are being developed to target the P2Y₂ receptor as well as the P2Y₄ and P2Y₆ receptors to stimulate Cl and fluid secretion in CF airways (77, 80, 126, 134, 135). Hwang and colleagues (61) demonstrated that the addition of purinergic agonists to the apical or basolateral membrane of rat primary tracheal epithelial cell cultures stimulates Cl secretion as a downstream result of activating purinergic receptors. This study also concluded that the P2Y receptors stimulated were different on the apical side vs. the basolateral side. Paradiso and colleagues (104) showed that purinergic agonists increase Ca when added to either side of an airway epithelium. Iwase et al. (67) also showed stimulation of Cl secretion in rabbit tracheal epithelium, whereas they also observed that ATP suppressed sodium absorption. This result was confirmed by Devor and Pilewski (38), who showed that UTP and ATP inhibited sodium absorption in human non-CF and CF airway epithelium. Taylor et al. (143) described how P2X receptor activation stimulated Cl secretion across the nasal epithelium of anesthetized mice, as well as across mouse and human primary epithelial cell monolayers. In IMCD monolayers, multiple P2Y and P2X receptors have been found (92). In the MDCK kidney epithelial model, Insel (63) found multiple P2Y receptors (P2X receptor expression in MDCK models has not been examined). Luo and colleagues (89) also found that multiple P2Y and P2X receptors were expressed on the luminal and serosal membranes of pancreatic duct, where they stimulated Cl secretion and increased Ca. Taken together, no matter what epithelium is studied, the same theme of multiple P2Y and P2X receptors being expressed in a given epithelial model rings true.

Collectively, these studies also describe ATP stimulation of Cl secretion and inhibition of sodium absorption across airway and kidney epithelia. Because Cl