Extracellular diadenosine polyphosphates play important signaling functions in a number of physiological responses. Here we show that diadenosine polyphosphates are normal constituents of tear fluid and are potent stimulators of tear secretion through their interaction with P2Y receptors. Diadenosine tetraphosphate (Ap4A) and Ap5A were found in rabbit tears under basal conditions at concentrations of 2.92 and 0.58 μM, respectively. Single applications of UTP, ATP, and Ap4A increased tear secretion to 160 ± 8% (n = 16) (P < 0.001), 131 ± 6% (P < 0.05), and 162 ± 11% (P < 0.05) of placebo values, respectively. Ap4A, Ap5A, and Ap6A, but not Ap2A and Ap3A, were able to stimulate tear secretion in a dose-dependent manner. Concentration-response studies produced pD2 values of 5.56 ± 0.03, 5.75 ± 0.12, and 5.50 ± 0.09 for Ap4A, Ap5A, and Ap6A, respectively, with Ap4A showing the greatest efficacy. Diadenosine polyphosphates also stimulated P2Y1 and P2Y2 receptors expressed in 1321N1 cells with no apparent effect on the other P2Y receptors tested. Nonselective P2 antagonists did not modify the tear secretion induced by UTP or Ap4A in rabbit eyes in vivo or in cloned receptors, except for a weak but significant reduction in stimulated tear secretion by reactive blue 2. These results suggest that diadenosine polyphosphates stimulate tear secretion via a P2Y receptor-mediated mechanism. Comparing the effects of diadenosine polyphosphates applied to the rabbit eye and to cloned P2Y receptors, it appears that the P2Y2 receptor subtype is responsible for the prosecretory effects of these compounds.
The discovery of diadenosine 5′-polyphosphates (ApnA,n = 2–7) (Fig. 1) and their release from platelets and chromaffin cells has led to many studies of the biological activity and cellular processing of these intra- and extracellular signaling molecules (Pintor, 1999; Hoyle et al., 2001). Diadenosine polyphosphates have interesting pharmacological effects on nucleotide receptors; that is, depending on the chain length, they may be agonists or antagonists at P2X and P2Y receptors with varying selectivity. There is not a clear relationship between the phosphate chain length and the selective activity on P2X or P2Y receptors; however, from a physiological point of view, some of them can act as vasodilators (Ap2A and Ap3A), whereas others act as vasoconstrictors (Ap4A, Ap5A, and Ap6A) (Ralevic and Burnstock, 1998).
Interest in uridine-containing nucleotides as extracellular signaling molecules increased with the discovery that UTP was a full agonist at the P2Y2 receptor, with potency comparable to that of the purine agonist ATP. Diadenosine polyphosphates were also found to be potent and full agonists in cells overexpressing the human P2Y2 receptor (Lazarowski et al., 1995). Also, the avian P2Y1 receptor is sensitive to Ap4A, presenting EC50values in the nanomolar range (Pintor et al., 1996). Important differences have been observed on native P2Y1receptors, where diadenosine tetraphosphate behaved as an antagonist in clear contrast to the behavior of this dinucleotide in cloned P2Y1 receptors (Vigne et al., 2000). On the other hand, other expressed P2Y receptors, such as P2Y4, are also activated by Ap3A-Ap6A in the micromolar range (Communi et al., 1996; Janssens et al., 1997). Recently, several of the pyrimidine dinucleotides UpnU (n = 2–7) have been prepared and their ability to stimulate P2Y receptors has been studied. Up4U is the most potent P2Y2 receptor agonist in the series (Pendergast et al., 2001). P2Y2 agonists are known to have prosecretory effects on the surface of the eye, such as stimulating the secretion of ions, fluid, and mucin (Hosoya et al., 1999; Fujihara et al., 2000, 2001; Li et al., 2001).
Despite various studies on P2Y2 agonists in ocular surface tissues, little is known about the role of ApnA signaling molecules on ocular physiology. Dinucleotides have been shown to affect intraocular pressure in rabbits: Ap4A lowered intraocular pressure, whereas others raised it (Peral et al., 2000). Here we report for the first time the presence of diadenosine polyphosphates in rabbit tears and suggest that these dinucleotides, through interactions with the P2Y2 receptor, are potent stimulators of tear secretion.
UTP and ATP were purchased from Amersham Biosciences, Inc. (Piscataway, NJ); UDP, Ap2A, Ap3A, Ap4A, Ap5A, Ap6A, and phosphodiesterase were purchased from Sigma Chemical (St. Louis, MO); 2MeSADP, pyridoxalphosphate-6-azophenyl-2′,4′-disulfonic acid (PPADS), suramin, and reactive blue-2 (RB-2) were purchased from Sigma/RBI (Natick, MA). Adrenoceptor antagonists (yohimbine and ICI 118,551) and cholinoceptor antagonists (hexamethonium and atropine) were obtained from Tocris (Bristol, UK). The purity of all nucleotide agonists was established by HPLC (95–99% purity). Schirmer strips were kindly provided by Allergan (Irvine, CA). Fluo-3/AM was obtained from Molecular Probes (Eugene, OR). Dulbecco's modified Eagle's medium, fetal bovine serum, G-418, and other cell culture reagents were obtained from the Tissue Culture Facility at the University of North Carolina (Chapel Hill, NC) or from Invitrogen (Carlsbad, CA). 1321N1 human astrocytoma cells stably expressing the P2Y1, P2Y2, P2Y4, P2Y6, or P2Y11 receptors, and wild-type 1321N1 cells were obtained from the University of North Carolina (Chapel Hill, NC).
Male New Zealand White rabbits weighing 2.0 to 2.5 kg were placed in individual cages with free access to food and water and subjected to regular cycles of light/dark (12 h). All the experiments were performed according to Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research and to the European Directive 86/609/EEC.
Tear Collection and Extraction.
The dinucleotide tear components were extracted from tears using Whatman no. 51 paper strips (Schirmer strips; Whatman, Maidstone, UK) placed in the inferior lid margin of the eye for 5 min. The strips were then placed in Eppendorf tubes containing 500 μl of ultrapure water and strongly vortexed for 5 min. The Schirmer strips were carefully rinsed and the liquid in the tube was heated in a 100°C bath for 20 min. Precipitated proteins were removed by centrifugation at 4000 rpm for 30 min. Diadenosine polyphosphates are stable to this treatment as previously reported; however, any mononucleotides present in the sample are destroyed by this sample workup procedure (Pintor et al., 1992). One hundred microliters of the supernatant was injected into the HPLC apparatus for analysis.
Identification and quantification of adenine dinucleotides in rabbit tears were performed by HPLC as previously described (Pintor et al., 1992). Briefly, the chromatographic system consisted of a Waters Novapak C18 column (15 cm in length, 0.4 cm in diameter), a 1515 Isocratic HPLC pump, a 2487 dual absorbance detector, and a Reodyne injector, all managed by the software Breeze from Waters (Milford, MA). The system was equilibrated overnight with 0.1 M KH2PO4, 3% methanol, pH 6.0, and detection of nucleotides was performed under isocratic conditions with the mobile phase described above at a flow rate of 1.5 ml/min. After injection of 100 μl of sample, detection was monitored at 260 nm. Peaks identified as putative diadenosine polyphosphates, based on comparing their retention times with the ones of commercial standards, were collected and subjected to phosphodiesterase treatment. Phosphodiesterase from Crotalus durissus (EC 126.96.36.199) at a concentration of 0.3 U/ml was incubated for 10 min with the corresponding putative dinucleotide, and the digestion products were analyzed by HPLC under the same conditions described previously. This treatment resulted in the quantitative hydrolysis of each compound isolated from tears and resulted in the appearance of two new peaks identified as adenine mononucleotides. Quantification of the products of hydrolysis of the phosphodiesterase was performed by comparing the areas under the curves with those of known amounts of commercial standards.
Measurement of Tear Secretion.
Tear secretion was measured according to the Schirmer test. Briefly, 10 μl of the test compound at the indicated concentrations was instilled via pipette in the eye. Thirty seconds later, a Schirmer strip was placed in the inferior lid margin of the eye for 5 min. Control experiments were performed by applying 10 μl of saline solution (0.9% NaCl). Tear secretion was measured as the length (millimeters) of the strip wetted by the tears.
Single-dose experiments were carried out by applying 10 μl of the corresponding nucleotide or dinucleotide at a concentration of 10 μg/μl. Dose-response analysis was performed by instilling doses ranging from 10−10 to 10−4 g/μl, always in a volume of 10 μl. Concentration-response curves were done by applying different doses in a noncumulative manner in one of the rabbit eyes, with the contralateral eye receiving the same volume of saline solution (control). Transformation of grams per microliter units into molar concentrations was performed by factoring in the corresponding molecular weight of each dinucleotide for each data point. Antagonists of P2 receptors such as PPADS, suramin, or RB-2 were applied 5, 10, 15, and 30 min before the application of any agonist. Adrenoceptor and cholinoceptor antagonists, yohimbine, ICI 118,551, atropine, and hexamethonium were applied at a concentration of 10 μg/μl (final volume 10 μl), 5 min before the instillation of agonist.
The values presented are the means ± S.E.M. of 8 to 12 experiments performed in 36 different animals. Statistical significance between treated and nontreated animals was estimated by the Student'st test.
1321N1 human astrocytoma cells stably expressing the human P2Y1, P2Y2, P2Y4, P2Y6, and P2Y11 receptors were grown in Dulbecco's modified Eagle's medium containing 4.5 g/l glucose, 5% fetal bovine serum, and 600 μg/ml G-418. For intracellular Ca2+ measurements, cells were seeded in 96-well black wall/clear bottom culture plates (3904; Corning Glassworks, Corning, NY), at a density of 35,000 cells/well and assays were conducted 2 days later when the cells had reached confluence.
Intracellular Ca2+ Measurements.
On the day of the assay, the growth medium in the culture plates was aspirated and replaced with 2.5 μM Fluo-3/AM in a final volume of 50 μl and incubated for 1 h at 25°C. Then the dye was replaced with assay buffer (10 mM KCl, 118 mM NaCl, 2.5 mM CaCl2, 1 mM MgCl2, 10 mM glucose, and 20 mM HEPES, pH 7.4) by using a Columbus Plate Washer (Tecan Inc., Research Triangle Park, NC). Intracellular Ca2+ levels in response to P2Y receptor agonists was monitored as changes in fluorescence intensity using a fluorescent light imaging plate reader (Pendergast et al., 2001) from Molecular Devices (Sunnyvale, CA). Average fluorescence units corresponding to peak height were captured on disk and exported for further analysis. Changes in fluorescence data corresponding to concentrations of intracellular Ca2+ were normalized to the response of the cognate agonists (2MeSADP for P2Y1 receptor, ATP for P2Y2receptor, UTP for P2Y4 receptor, UDP for P2Y6 receptor, and ATP for P2Y11 receptor).
Agonist potencies were calculated using a four-parameter logistic equation and the GraphPad software package (GraphPad, San Diego, CA). EC50 values (mean ± standard error) represent the concentration of agonist at which 50% of the maximal effect is achieved. Three experiments with triplicate assays were conducted on separate days for each P2Y receptor subtype.
Presence of Diadenosine Polyphosphates in Rabbit Tears.
Diadenosine polyphosphates were isolated from rabbit tears and quantified by HPLC as indicated under Experimental Procedures. Chromatographic separation of tear extracts resulted in two well defined peaks that were identified as Ap4A and Ap5A by comparing their retention times with those of authentic external standards (Fig.2, A and B). To confirm the identity of these two dinucleotides, samples were rechromatographed and enriched with 25 pmol of commercial Ap4A and Ap5A. The subsequent chromatographic analysis demonstrated the coelution of the standards and the peaks found in the sample, thus suggesting that both naturally occurring peaks and commercial diadenosine polyphosphates are the same (data not shown).
The identity of these compounds was fully confirmed by rechromatography of these peaks after C. durissusphosphodiesterase treatment. This enzyme cleaves dinucleotides forming AMP plus another mononucleotide whose phosphate length depends on the original dinucleotide. Enzymatic digestion of the peak identified as Ap4A resulted in the formation of AMP and ATP, confirming its identity as diadenosine tetraphosphate (Fig.3A). When the same experimental protocol was applied to the Ap5A peak, the products of the enzymatic digestion obtained were AMP and adenosine tetraphosphate (Fig. 3B). These results clearly demonstrated that Ap4A and Ap5A are normal constituents of rabbit tears. The concentrations of adenine dinucleotides found in rabbit tears were 2.92 ± 0.28 μM for Ap4A and 0.58 ± 0.11 μM Ap5A (n = 8).
Effect of Mono- and Dinucleotides on Tear Secretion.
To determine whether mononucleotides were able to modify rabbit tear secretion, single doses of ATP, UTP, ADP, and UDP were applied at 10 μg/μl (final volume 10 μl). As shown in Fig.4A, among the tested mononucleotides, UTP and ATP significantly increased Schirmer scores to 160 ± 8% (n = 16) (P < 0.001) and 131 ± 6%, respectively (n = 12), compared with control. UDP and ADP, on the other hand, did not significantly increase tear secretion, their values being 105 ± 2% for UDP and 107 ± 1% for ADP (n = 10).
When diadenosine polyphosphates were assayed under the same conditions as the mononucleotides, Ap4A, Ap5A, and Ap6A significantly increased tear secretion by 162 ± 3, 126 ± 6, and 125 ± 15%, respectively (P < 0.05,n = 12). Neither Ap2A nor Ap3A was able to significantly change tear secretion rates (95 ± 2 and 89 ± 7%, respectively,n = 10) (Fig. 4B).
Concentration-effect curves of all the dinucleotides in the range of 10−10 to 10−4 g/μl showed pD2 values for Ap4A, Ap5A, and Ap6A of 5.56 ± 0.03, 5.75 ± 0.12, and 5.50 ± 0.09 (n = 8). These values corresponded to EC50 values of 2.76 μM for Ap4A, 1.77 μM for Ap5A, and 3.16 μM for Ap6A. Ap4A was the dinucleotide eliciting the strongest effect, 162 ± 2.4%, with Ap5A and Ap6A exhibiting similar maximal effects (125 ± 7%) (Fig.5). Diadenosine diphosphate and diadenosine triphosphate failed to produce any change on tear secretion even at the highest concentrations assayed (n = 8).
Effects of Antagonists on Tear Secretion.
The activity of diadenosine polyphosphates on tear secretion suggests the activation of P2 nucleotide receptors. To confirm this, the effects of three nonselective, nucleotide receptor antagonists on tear secretion were studied: PPADS, suramin, and RB-2. The pretreatment with these antagonists for 5 min before the administration of nucleotides did not modify tear secretion induced by either Ap4A or UTP (10 μg/μl, n = 10; data not shown). Only RB-2 modified very slightly the Schirmer scores of both compounds (n = 12), reducing tear secretion as observed in Fig.6. Similar effects were observed when the rabbits were pretreated with antagonists for 10, 15, or 30 min. In rabbits, RB-2 reduced tear secretion from 155 to 132% for UTP and from 150 to 125% for Ap4A. Although PPADS and suramin have been shown to antagonize the P2Y1 receptor, their effects on other P2Y receptor subtypes are less known (Charlton et al., 1996; Schachter et al., 1996). In our hands, addition of varying concentrations of the nonselective P2 receptor antagonists PPADS, suramin, and RB-2 to cells transfected with P2Y2, P2Y4, or P2Y6 receptors revealed no meaningful antagonism of the calcium responses (data not shown).
Because diadenosine polyphosphates are known to be stored together with classical neurotransmitters, such as catecholamines and acetylcholine, we investigated whether basal and dinucleotide-induced tear secretion are influenced by adrenergic and cholinergic blockade. Blockade of muscarinic and nicotinic receptors with 10 μg/μl (10 μl) atropine and 10 μg/μl (10 μl) hexamethonium produced a decrease in basal tear secretion by 51.3%, whereas the blockade of α1- and α2-adrenoceptors decreased basal tear secretion by 18% (Fig. 7). In contrast, adrenoceptor and cholinoceptor antagonists did not affect tear secretion evoked by application of Ap4A or UTP (n = 10).
Pharmacological Profile of Diadenosine Polyphosphates on Transfected P2Y Receptors.
Adenine dinucleotides were tested for their ability to activate human P2Y1, P2Y2, P2Y4, P2Y6, and P2Y11 receptors as measured by the mobilization of intracellular Ca2+. P2Y4, P2Y6, and P2Y11 were insensitive to diadenosine polyphosphates and were only stimulated by different mononucleotides (data not shown). Two receptors, P2Y1 and P2Y2, were fully activated by diadenosine polyphosphates with different pharmacological patterns (Fig. 8, A and B). The P2Y1 and P2Y2 receptors presented EC50 and corresponding pD2 values as shown in Table1. The P2Y1receptor was fully activated by submicromolar concentrations of Ap3A, whereas the P2Y2receptor was activated by Ap4A at this low concentration. The diadenosine polyphosphates were all inactive at the other P2Y receptors, and, as stated above, no meaningful antagonism of the P2Y receptors was observed with PPADS, suramin, or RB-2 (data not shown).
The results show for the first time that New Zealand White rabbit tears contain diadenosine polyphosphates and that these compounds have effects on tear secretion. The concentrations of diadenosine polyphosphates in rabbit tears, as demonstrated by HPLC, were in the micromolar range, which is sufficient to allow them to activate nucleotide receptors present on the ocular surface. Applied dinucleotides altered the rate of tear production: dinucleotides with four, five, and six phosphates increased tear secretion, whereas Ap2A and Ap3A were inactive. Ap4A was the most efficacious at increasing tear secretion, giving values similar to that obtained with UTP.
Diadenosine di- and triphosphates were unable to change tear secretion even when they were assayed at the highest concentrations. The concentrations necessary to increase tear secretion (EC50 for Ap4A, 2.27 μM) correspond well with the basal concentration of Ap4A determined under normal conditions (2.92 μM). This indicates that Ap4A is in tears at concentrations high enough to stimulate tear secretion, and therefore it may be a naturally occurring compound that promotes a normal tear secretion rate.
Regarding the nucleotide receptor subtype that may be involved in this physiological action, the lack of effect of UDP (P2Y6 agonist) and ADP (P2Y1 agonist), and, moreover, the full effect of Ap4A, strongly suggested the involvement of a P2Y2 receptor (Ralevic and Burnstock, 1998). This is consistent with P2Y receptors expressed in 1321N1 human astrocytoma cells. Among the receptors studied, only P2Y1 and P2Y2 were able to generate changes in the intracellular Ca2+ when challenged with diadenosine polyphosphates. When the agonist profile of both P2Y1 and P2Y2 receptors expressed in astrocytoma cells is compared with that obtained for tear secretion in the rabbits, it seems clear that the rabbit P2Y receptor matches better with the P2Y2 receptor than with the P2Y1 receptor. Although the expressed P2Y4 receptor was not activated by diadenosine polyphosphates, previous studies have shown that in addition to being activated by UTP, UDP, and ATP, it is also activated by ApnA (Communi et al., 1996; Bogdanov et al., 1998; Kennedy et al., 2000; Patel et al., 2000). However, the activity of ATP is very low and there is no discrimination among the diadenosine polyphosphates (Communi et al., 1996). In the rabbit eye, only RB-2 was able to partially inhibit the induction of tear secretion by Ap4A or UTP.
Blockade of adrenoceptors and cholinoceptors with the use of the corresponding antagonists did not alter the effect produced by adenine dinucleotides, in contrast to the partial reduction in basal tear secretion, as has been previously described (Salvatore et al., 1999). These results are consistent with the observation that P2Y2 receptor gene expression is observed throughout the corneal and conjunctival epithelium, but not found in nerve terminals innervating the meibomian gland (Yerxa et al., 2000). Thus, the dinucleotides are likely acting at postjunctional rather than prejunctional sites.
Agonists of P2Y2 receptors such as dinucleoside polyphosphates may prove to be useful in the development of compounds for ameliorating conditions in which enhanced tear secretion would be beneficial, such as dry eye. INS365, a diuridine polyphosphate P2Y2 receptor agonist, is more efficacious at stimulating tear secretion than Ap4A, showing a maximal 2-fold increase in tear secretion in the same rabbit model versus saline control (Yerxa et al., 1999). Positive safety and efficacy results of INS365 in dry eye patients were recently reported, and this compound is currently undergoing definitive evaluation in phase III clinical trials (Foulks et al., 2001). Additionally, ApnAs appear to be compounds with potential therapeutic value as in the case of ocular hypertension (Peral et al., 2000).
The presence of diadenosine polyphosphates in tears and their effect in stimulating tear secretion introduce new physiological elements in the regulation of tear secretion. The control of tear secretion is partially performed by the nervous system. Although it has been demonstrated that denervation may partially affect tear secretion, it has been shown that this process does not suppress tear secretion (Meneray et al., 1998). An explanation for this may be the presence of diadenosine polyphosphates such as Ap4A, which would stimulate tear secretion in the absence of neural inputs. This would indicate that diadenosine polyphosphates, as occurs with ATP, may be released upon mechanical stimulation of corneal epithelial cells (Jensen, 2000; Srinivas and Fleiszig, 2000).
Activation of P2Y2 receptors by diadenosine polyphosphates appears to be one of the chief actions of these signaling molecules, causing an increase in intracellular calcium, chloride, and fluid secretion; mucin secretion; and activation of downstream signaling events via the mitogen-activated protein kinase cascade (Fujita et al., 2001). With the isolation and characterization of nucleotides on the ocular surface now more clearly defined, it is possible to begin to construct an overall scheme of nucleotide release, action, metabolism, and reuptake. Nucleotides and dinucleotides are released onto the ocular surface in concentrations relevant to activate P2X and P2Y receptors. Their subsequent metabolism by ectonucleotidases (Gukasyan et al., 2002) can then generate different P2 receptor ligands with further physiological actions, although the presence and roles of these other receptors on the ocular surface remain to be elucidated. Finally, the resulting nucleoside base, e.g., adenosine, is then reabsorbed via nucleoside transporters (Hosoya et al., 1998). Although more experiments are necessary to confirm this hypothesis, nucleotides, dinucleotides, and their associated receptors and proteins are important to ocular surface signaling.
In summary, we present here evidence that Ap4A and Ap5A are present in rabbit tears and that exogenous application of diadenosine polyphosphates can facilitate tear secretion. This effect seems to be mediated by P2Y2 receptors, according to the pharmacology obtained in rabbits when the effect of diadenosine polyphosphates is compared with expressed P2Y receptors. These findings suggest that these important signaling molecules may be involved in the regulation of basal tear secretion. Moreover, stimulation of this alternate, nonadrenergic, noncholinergic secretion mechanism may be useful for the treatment of dry eye syndrome.
We thank Wendy Anders for help with manuscript preparation and José Boyer for helpful comments.
This study was funded by Inspire Pharmaceuticals, Inc. (Durham, NC). Portions of this report were presented at the annual meeting of the Association for Research in Vision and Ophthalmology, Fort Lauderdale, FL, 2001 April 29–May 4 (Pintor JJ, Peral A, Hoyle CHV, Redick K, Douglass J, Sims I, and Yerxa B (2001)Invest Ophthomol Vis Sci 42:S261).
- diadenosine polyphosphate
- pyridoxalphosphate-6-azophenyl-2′,4′-disulfonic acid
- reactive blue 2
- high-performance liquid chromatography
- ICI 118,551
- Received August 25, 2001.
- Accepted September 28, 2001.
- The American Society for Pharmacology and Experimental Therapeutics