Dinucleotides as Growth-promoting Extracellular Mediators

Dinucleoside diphosphates, Ap2A, Ap2G, and Gp2G represent a new class of growth-promoting extracellular mediators, which are released from granules after activation of platelets. The presence of theses substances was shown after purification from a platelet concentrate. The substances were identified by UV spectrometry, retention time comparison with authentic substances, matrix-assisted laser desorption/ionization mass spectrometry, post-source-decay matrix-assisted laser desorption/ionization mass spectrometry, and enzymatic analysis. Ap2A, Ap2G, and Gp2G have growth-stimulating effects on vascular smooth muscle cells in nanomolar concentrations as shown by [3H]thymidine incorporation measurements. The calculated EC50 (log m; mean ± S.E.) values were −6.07 ± 0.14 for Ap2A, −6.27 ± 0.25 for Ap2G, and −6.91 ± 0.44 for Gp2G. At least 61.5 ± 4.3% of the dinucleoside polyphosphates are released by platelet activation. The intraplatelet concentrations suggest that, in the close environment of a platelet thrombus, similar dinucleoside polyphosphate concentrations can be found as in platelets. Intraplatelet concentration can be estimated in the range of 1/20 to 1/100 of the concentration of ATP. In conclusion, Ap2A, Ap2G, and Gp2G derived from releasable granules of human platelets may play a regulatory role in vascular smooth muscle growth as growth-promoting mediators.

To further evaluate the pathogenesis of hypertension, there is a continued interest in the identification of novel endogenous compounds with growth-stimulating effects on vascular smooth muscle cells (VSMCs). 1 In this context, novel endogenous nu-cleotides have been recognized as powerful vasoactive messengers.
From these results the question arose as to whether P 1 ,P 2dinucleoside diphosphates containing two adenosine or guanosine groups or an adenosine and a guanosine group also occur in humans. There is one report on the existence of diadenosine diphosphate (Ap 2 A) isolated from human cardiac tissue (12). In contrast to Ap 2 A the P 1 ,P 2 -dinucleoside diphosphates Ap 2 G and Gp 2 G have not been described as endogenous substances so far in the literature.
Here the existence of diadenosine diphosphate (Ap 2 A), adenosine guanosine diphosphate (Ap 2 G), as well as diguanosine diphosphate (Gp 2 G) in releasable granules of human platelets is shown for the first time and their growth-stimulating effect on cultured vascular smooth muscle cells is described.

Chemicals
HPLC water (gradient grade) and acetonitrile were from Merck (Germany). All other substances were purchased from Sigma (Germany). * This study was supported by a grant of the Deutsche Forschungsgemeinschaft (DFG: Schl 406/2-1). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ʈ

Purification of Dinucleoside Diphosphates from Human Platelets
Dinucleoside diphosphates were isolated from human platelets unsuitable for transfusion. The platelets were suspended in an isotonic salt solution and centrifuged at 2500 ϫ g for 5 min. The pellet was resuspended in an isotonic salt solution and centrifuged again (2500 ϫ g, 5 min). The supernatant was aspirated, and the platelet pellet was frozen to Ϫ30°C and rethawed in bidistilled water (step 1). Then the resulting suspension was deproteinized (step 2) with 0.6 M perchloric acid (final concentration). After adjusting the pH to 7.0 with 5 M KOH, the precipitated protein and KClO 4 were removed by centrifugation.

Chromatography
Preparative Reversed Phase Chromatography-Triethylammonium acetate (TEAA) was added to the supernatant (final TEAA concentration 40 mM), and hydrophobic substances were concentrated on a C18 reversed phase column (Lichroprep, 310 ϫ 65 mm, 40 -65 m, Merck, Germany) using 40 mM TEAA in water (eluent A; flow, 2 ml/min). After removing substances not binding to the column with aqueous 40 mM TEAA (flow: 2 ml/min), the adsorbed substances were eluted with 20% acetonitrile in water (eluent B). The elution was detected by measuring the UV absorption at 254 nm. The eluate was lyophilized and stored frozen at Ϫ80°C (step 3).
Affinity Chromatography-The lyophilized eluate of the reversed phase chromatography was dissolved in 1 M ammonium acetate (eluent C; pH 9.5) and purified further with affinity chromatography (step 4). The affinity chromatography gel, phenyl boronic acid coupled to a cation-exchange resin (BioRex 70, Bio-Rad), was synthesized according to Barnes (13). The affinity resin was packed into a glass column (150 ϫ 20 mm) and equilibrated with 1 M NH 4 Ac (pH 9.5; flow, 2 ml/min). The pH of the eluate from the reversed phase column was adjusted to pH 9.5 and loaded to the affinity column. The column was washed with 1 M NH 4 Ac (pH 9.5) with a flow rate of 2 ml/min. Binding substances were eluted with 1 mM HCl (eluent D). Fractions were monitored with a UV detector at 254 nm. The eluate was frozen and lyophilized.
Reversed Phase Chromatography-The eluate of the affinity chromatography was desalted by reversed phase chromatography (step 4). The reversed phase column (Supersphere, 210 ϫ 4.1 mm, 4 m, Merck) was equilibrated with aqueous 40 mM TEAA solution (eluent A). The sample, with 40 mM TEAA in water added, was pumped at a rate of 0.5 ml/min onto the column. After washing the column with 15 ml of eluent A, the fraction of interest was eluted with 35% acetonitrile in water (eluent E). The resulting fractions were lyophilized and stored at Ϫ80°C.

Matrix Assisted Laser Desorption/Ionization Mass Spectrometry
The desalted and lyophilized fractions of the anion-exchange chromatography were examined by matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) (14). A reflectron-type time-offlight mass spectrometer (Reflex III, Bruker-Franzen, Germany) was used according to Hillenkamp (14). The sample was mounted on an x, y, z movable stage allowing for irradiation of selected sample areas. In this study, a nitrogen laser (Laser Science Inc.) with an emission wavelength of 337 nm and 3-ns pulse duration was used. The laser beam was focused to a diameter typically of 50 m at an angle of 45°to the surface of the target. Microscopic sample observation was possible via a diachronic mirror in the beam path. 10 -20 single spectra were accumulated for a better signal-to-noise ratio. The concentrations of the analyzed substances were 1-10 M in double-distilled water. 1 l of the analyte solution was mixed with 1 l of the matrix solution. For this study, a solution of 50 mg/ml 3-hydroxypicolinic acid was used. For calibration of the mass spectra, diadenosine hexaphosphate (Ap 6 A) was used as external standard. The mixture was gently dried on an inert metal surface before introduction into the mass spectrometer. The mass accuracy was in the range of ϳ0.05%.

UV Spectroscopy
The desalted, lyophilized fractions of the reversed phase chromatography (step 8) were dissolved in water (100 l). To measure UV spectra at different pH the pH values of the solutions were adjusted to 3.0, 7.0, and 9.0 by 0.1 M HCl and 0.1 M NaOH, respectively. The UV absorbance of the fractions were determined by a UV-visible spectrometer (DU-600, Beckman) at wavelengths between 190 and 400 nm with a scan speed of 400 nm/min.

Platelet Activation by Thrombin and Purification of Dinucleoside Diphosphates AP 2 A, AP 2 G, GP 2 G, and Serotonin from the Supernatant
Three platelet concentrates (each 200 ml; 10 7 platelets/l) were resuspended in 600 ml of a buffer containing 0.14 M NaCl, 0.15 mM Tris-HCl. To prevent premature activation 0.35% (w/v) albumin was added to the buffer. The resuspended platelet concentrates were divided into three parts.
To test the release of the dinucleoside diphosphates, one aliquot was incubated with thrombin (0.05 units/ml) for 1 min. Preliminary experiments showed that fibrinogen binding in platelets did not exceed 2000 molecules/cell. After stimulation with thrombin, the fibrinogen binding rose 20-to 30-fold. Determination of fibrinogen was performed exactly as described previously (15). Moreover, the concentration of serotonin was determined in the supernatant. As control the second aliquot was not incubated with thrombin.
For purification of dinucleoside diphosphates Ap 2 A, Ap 2 G, and Gp 2 G from the supernatant, platelets were removed by centrifugation (4000 rpm, 4°C, 10 min). The supernatant was deproteinated with 0.6 M (final concentration) perchloric acid and centrifuged (4000 rpm, 4°C, 5 min). After adjusting the pH to 7.0 with 5 M KOH the precipitated proteins and KClO 4 were removed by centrifugation (4000 rpm, 4°C, 5 min). The supernatants of both aliquots of the platelet concentrates were chromatographed according to chromatographic steps for the purification of dinucleoside diphosphates from platelets. Dinucleoside diphosphates Ap 2 A, Ap 2 G, and Gp 2 G were identified by retention time comparison with authentic substances as well as MALDI-MS.
For the measurement of the total endogenous serotonin content, a method described by Hervig et al. (16) was used. Briefly, 600 l of the platelet concentrate as prepared above was mixed with 200 l of a 2.8 M perchloric acid solution containing dithiothreitol (40 mM) to precipitate the proteins. The precipitate was removed by centrifugation (8000 ϫ g, 2 min), and 520 l of the supernatant was neutralized with 130 l of 3 M K 2 HPO 4 . The precipitated potassium perchlorate was removed by a second centrifugation (8000 ϫ g, 2 min).
The supernatant was transferred and was directly analyzed by the reversed phase chromatographic method of Anderson et al. (17). 100 l of the supernatant was injected onto a reversed phase column (Supersphere, 210 ϫ 4.1 mm, 4 m, Merck) eluted with 0.1 M phosphate buffer (pH 4.5) containing 250 l/liter triethylamine, 150 mg/liter sodium octylsulfate, and 20% (v/v) methanol (flow rate, 0.5 ml/min). The fluorescence was detected using an SP920 intelligent fluorescence detector (Jasco) with excitation and emission wavelength settings of 285 and 350 nm, respectively. Quantification of serotonin was done by using a calibration curve.
For the measurement of the released serotonin after thrombin stimulation, a method described by Hervig et al. (16) was used. Briefly, 600 l of the platelet concentrate as prepared above were incubated with 10 NIH units of thrombin (10 l) for 10 min. After removing the platelet remnant by centrifugation (8000 ϫ g, 30 s), 450 l of supernatant was mixed with 150 l of the perchloric acid/dithiothreitol solution and centrifuged again (8000 ϫ g, 2 min). 400 l of the supernatant was neutralized with 100 l of a 3 M solution of K 2 HPO 4 , recentrifuged as above, and injected to the chromatography using the method described by Anderson et al. (17).

Synthesis and Chromatography of Authentic P 1 ,P 2 -dinucleoside Diphosphates
In contrast to diadenosine diphosphate and diguanosine diphosphate, adenosine guanosine diphosphate was commercially not available. Therefore, synthesis of adenosine guanosine diphosphate was necessary to control the authenticity of the isolated substances. Adenosine guanosine diphosphate was synthesized and chromatographed following a study described elsewhere (18). Briefly, Ap 2 G was synthesized by mixing AMP (25 mM) and GMP (25 mM) as substrates in the presence of 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (2.5 M), HEPES (2 M), and magnesium chloride (125 mM). The substances were dissolved in water, thoroughly mixed with a vortex mixer, and incubated at 37°C at pH 6.5 for 48 h. The reaction mixture was concentrated on a preparative C18 reversed phase column (condition described above). The concentrate was displaced on a reversed phase column (carrier: 40 mM TEAA in water (eluent A); displacer: 160 mM n-butanol (eluent K), flow 100 l/min). As a result of displacement chromatography, anion-exchange chromatography yielded baseline separated dinucleoside diphosphates (Ap 2 A, Ap 2 G, and Gp 2 G).
Commercially available diadenosine diphosphate and diguanosine diphosphate are contaminated with mononucleotides. Therefore, these P 1 ,P 2 -dinucleoside diphosphates were purified by displacement chromatography using a reversed phase column (conditions as describe above) before testing the authenticity of the isolated substances.

Enzymatic Cleavage Experiments
Aliquots of the fractions containing homogenous nucleotides from the reversed phase chromatography (steps 7 and 8 of the purification procedure), were incubated with enzymes as described in the following. The samples were dissolved: (a) in 20 l 200 mM Tris buffer (pH 8.9) and incubated with 5Ј-nucleotide hydrolase (3 milliunits (mU); from Crotalus durissus, EC 3.1.15.1, from Roche Molecular Biochemicals, Germany, purified according to Sulkowski and Laskowski (19)

Cell Proliferation Assay with Aortic Smooth Muscle Cells
Aortic smooth muscle cells (VSMCs) from normotensive Wistar-Kyoto rats were subcultured in 96-well dishes (Falcon) at a density of 5 ϫ 10 4 cells/ml and kept in culture medium containing 10% fetal calf serum (FCS) to reach a subconfluent monolayer. After 24 h, the cells were growth-arrested in 0.1% FCS for 48 h without affecting cell adherence to culture wells. Quiescent VSMCs were then exposed to fresh culture medium with 0.1% FCS with and without the tested agonists for another 48-h incubation period. Cell proliferation was measured using the [ 3 H]thymidine incorporation rate as described elsewhere (20). The viability of VSMCs was tested using trypan blue exclusion test. The viability was 95.2 Ϯ 3.5% under control conditions and 93 Ϯ 4.1% after stimulation with dinucleoside diphosphates.

Cell Proliferation Assay with Fibroblasts
Human skin fibroblasts were obtained from the Human Genetic Mutant Cell Repository Institute for Medical Research (Camden, NJ) and cultured over several passages after detachment of the confluent cells with Puck's Saline A physiological solution (21) containing 0.04% trypsin and 0.02% EDTA buffer. The cells were allowed to grow as described for VSMCs.
Fibroblasts were seeded in 24-well culture plates and grown to confluence. Then the medium was replaced by serum-free medium consisting of a mixture of Dulbecco's modified Eagle's medium and Ham's F-10 medium (1:1). Following another 24-h cultivation in serum-free medium, stimuli were added and cells were exposed to the stimulating agents for 20 h before 3 Ci/ml [ 3 H]thymidine was added to the serumfree medium. Four hours later, experiments were terminated by aspirating the medium and subjecting the cultures to sequential washes with phosphate-buffered saline containing 1 mM CaCl 2 , 1 mM MgCl 2 , 10% trichloroacetic acid, and ethanol/ether (2:1, v/v). Acid-insoluble [ 3 H]thymidine was extracted into 250-l dishes with 0.5 M NaOH, and 100 l of this solution was mixed with 5 ml of scintillant (Packard, Ultimagold) and quantified using a liquid scintillation counter (Beckman LS 3801, Dü sseldorf, Germany).

RESULTS
Human platelets were isolated (step 1) and deproteinated with perchloric acid (step 2), and the supernatant nucleotides were concentrated by ion-pair reversed phase chromatography (step 3). In the following steps, isolation and identification of dinucleoside diphosphates from human platelets is exemplified for Gp 2 G.
After mononucleotides were separated from dinucleotides by affinity chromatography (13) (step 4) the desalted and lyophilized eluate (step 5) was fractionated by anion-exchange chromatography (step 6). The anion-exchange chromatogram is shown in Fig. 1A. Although P 1 ,P 2 -dinucleoside diphosphates have the same charge, P 1 ,P 2 -dinucleoside diphosphates Ap 2 A, Ap 2 G, and Gp 2 G were separated because of hydrophobic interaction between the anion-exchanger and the P 1 ,P 2 -dinucleoside diphosphate.
Fractions of the anion-exchange chromatography with a significant absorbance at 254 nm were separated by reversed phase chromatography (step 7). In Fig. 1B the chromatogram of the reversed phase chromatography is given. The substance eluting at a retention time of 27 min was rechromatographed by reversed phase chromatography (step 8) using the same conditions as before (step 7).
In the last chromatographic step (step 8), a single UV peak was obtained (Fig. 1C). The substance underlying this peak was identified by the following results: (a) The substance chromatographed to homogeneity was analyzed by MALDI-PSD mass spectrometry revealing a molecular mass of 709.4 ( Fig.  2A). Each signal was assigned to a fragment of Gp 2 G as shown in Table I. The MALDI-PSD spectrum was completely identical to that of authentic Gp 2 G (14). (b) The UV spectrum of guanine was obtained from the rechromatographed substance, including the characteristic shift obtained by acidification to pH 3.0, 7.0, and 9.0 ( Fig. 2B; Table II) (22). (c) The retention time of the isolated fraction in step 8 was identical to that of authentic Gp 2 G (18). (d) Cleavage of the molecules with 5Ј-nucleotide hydrolase (from C. durissus) yielded GMP, as evidenced by MALDI mass spectra and by retention times identical with those of authentic Gp 2 G. The cleavage pattern was identical to that of synthetic Gp 2 G. Incubation of the molecule with 3Јnucleotide hydrolase (calf spleen) and alkaline phosphatase yielded no cleavage products. The enzymatic cleavage experiments demonstrate that the polyphosphate chain interconnects the guanosines via phosphoester bonds with the 5Ј-oxygens of the riboses (Fig. 2C).
In analogous manner also Ap 2 A as well as Ap 2 G were purified from human platelets and identified by the signal pattern of the PSD-MALDI-MS fragmentations (Table I), enzymatic cleavage experiments, and UV spectroscopy (Table II).
After isolation and identification of P 1 ,P 2 -dinucleoside diphosphates from human platelets, the question arose as to whether P 1 ,P 2 -dinucleoside diphosphates are released in the extracellular space. Fig. 4 shows the anion-exchange chromatograms of a platelet suspension (Fig. 4A) and a supernatant from an equivalent platelet suspension aggregated with thrombin (Fig. 4B). P 1 ,P 2dinucleoside diphosphates can be found in the supernatant after platelet aggregation (labeled in Fig. 4B by arrows) but not in the supernatant of unstimulated platelets. The intracellular amount of P 1 ,P 2 -dinucleoside diphosphates in intact human platelets can be estimated in the range of 0.5-2.0 attomol/platelet. From the concentrations determined in the supernatant, the portion released upon platelet aggregation was estimated as 61.5 Ϯ 4.3% for each P 1 ,P 2 -dinucleoside diphosphates. The intracellular amount of serotonin was 3.2 Ϯ 0.5 attomol/platelet. In the supernatant of unstimulated platelets serotonin was not detectable. After platelet stimulation with thrombin the serotonin amount of supernatant was 2.2 Ϯ 0.4 attomol/platelet, indicating that 68.7 Ϯ 12.6% of the intracellular serotonin amount was released by thrombin stimulation. The comparable degree of secretion of P 1 ,P 2 -dinucleoside diphosphates and serotonin suggests that both classes of agents are released in a quantitatively similar fashion. DISCUSSION The findings revealed that P 1 ,P 2 -dinucleoside diphosphates Ap 2 A, Ap 2 G, and Gp 2 G are endogenous messengers of human platelets.
The results of the cell proliferation assay show that Ap 2 A, Ap 2 G, and Gp 2 G act as potent growth mediators of VSMCs. The maximum effect of Ap 2 A, Ap 2 G, and Gp 2 G on VSMC proliferation rate was about one order of magnitude less, and the threshold concentration was about one order of magnitude higher than for PDGF, indicating that the dinucleoside phosphates are weaker growth factors than PDGF. Nevertheless, it has to be kept in mind that the local concentrations of these nucleotides after platelet aggregation probably are much higher than the physiological PDGF concentrations.
The receptor-mediating vascular growth is not yet known, although a P 2 purinoceptor subtype is most likely. Especially,  the P 2 Y 2 purinoceptor may be considered, because ATP and GTP binding to this receptor cause similar mitogenic effects in VSMCs (23). At present, the growth-stimulating effect of dinucleoside diphosphates is only demonstrable in VSMCs. The growth of fibroblasts is not affected by dinucleoside diphosphates. This result may represent a different expression of purinoceptors on VSMCs and fibroblasts. In contrast to Ap n G and Gp n G with n ϭ 3-6 (7), Ap 2 A, Ap 2 G, and Gp 2 G do not potentiate the growth-stimulating effect of PDGF. Presently it is open to speculation whether this different behavior reflects activation of different purine receptor subtypes.
Because the P 1 ,P 2 -dinucleoside diphosphates are released upon platelet activation, their growth-stimulating effect may contribute to that of PDGF and other growth mediators released from platelets. Therefore, together with known growth mediators, the described nucleotides may also participate in initiating atherosclerotic lesions.
Obviously, Ap 2 A, Ap 2 G, and Gp 2 G may exert their effects after release by platelet activation as is known for the diadenosine polyphosphates Ap n A (with n ϭ 3-6) (3, 24) and for the Ap n Gs and Gp n Gs with (n ϭ 3-6) (7).
From the intracellular amount of P 1 ,P 2 -dinucleoside diphosphates in intact human platelets, the intracellular concentration of P 1 ,P 2 -dinucleoside diphosphates in intact human platelets can be calculated as 0.1-0.4 mM (volume of a platelet: 5.2 fl (25)). In platelets, two pools of nucleotides have been demonstrated (26). One pool is utilized for the metabolic needs of the platelets. The second pool, the dense granules, is a storage pool, which can be released into the extracellular space. As demonstrated, serotonin and dinucleoside diphosphates Ap 2 A, Ap 2 G, and Gp 2 G are released in parallel, it can be assumed that dinucleoside diphosphates are costored in dense granula with serotonin. The concentration of P 1 ,P 2 -dinucleoside diphosphates in the dense granula can be estimated to be 0.2-0.8 mM, assuming that 50% of total volume of human platelets constitutes dense granula (27).
The extracellular dinucleoside polyphosphate concentrations occurring after platelet activation depend on the extracellular volume of distribution. The intraplatelet concentrations suggest that, in the close environment of a platelet thrombus, similar dinucleoside polyphosphate concentrations can be found as in platelets. Therefore, the maximum extracellular concentration of P 1 ,P 2 -dinucleoside diphosphates can be calculated as 0.2-0.8 mM in accordance to the concentration of P 1 ,P 2 -dinucleoside diphosphates in dense granula. The minimum concentration can be correspondingly estimated as 0.1 Ϫ 0.4 M in accordance with the concentration of P 1 ,P 2 -dinucleoside diphosphates after the release into the surrounding blood volume of the platelets. Theses estimations demonstrate that the extracellular concentrations of P 1 ,P 2 -dinucleoside diphosphates are sufficient for affecting the rate of proliferation of vascular smooth muscle cells.
How are these substances biosynthesized? The enzymes involved in synthesis of diadenosine polyphosphates are only partially known, and none of the known enzymes are described in human platelets (for review see Ref. 28). Aminoacyl-tRNA synthetases catalyze the formation of Ap 3 A and Ap 4 A (aminoacyl-AMP ϩ ADP 3 Ap 3 A, aminoacyl-AMP ϩ ATP 3 Ap 4 A) (29). Adenosine 5Ј-monophosphate does not react with this enzyme (30), and therefore this type of enzymatic reaction cannot yield Ap 2 A. Ap 4 A phosphorylases are another class of diadenosine polyphosphate-synthesizing enzymes according to the following reaction ADP ϩ ATP 3 Ap 4 A ϩ P i (29). Theoretically, the reaction of a diadenosine polyphosphate phosphorylase catalyzing the formation of Ap 2 A should be AMP ϩ ADP 3 Ap 2 A ϩ P i . Alternatively, a nonenzymatic synthesis may be considered. Given that mostly mononucleotides such as AMP are found together with biogenic amines such as catecholamines, the coexistence of both nucleotides and amines within the same subcellular localization may allow a nonenzymatic reaction generating diadenosine polyphosphates. From AMP and a biogenic amine a phosphoramidate may be generated, which is a highly reactive intermediate. A further reaction with another AMP could then yield diadenosine diphosphate (Ap 2 A). At present no definite answer can be given by which biochemical pathway P 1 ,P 2 -dinucleoside diphosphates are synthesized in human platelets.
In conclusion, releasable granules of human platelets contain diadenosine diphosphate (Ap 2 A), adenosine guanosine diphosphate (Ap 2 G), as well as diguanosine diphosphate (Gp 2 G), which are potent growth-stimulating mediators in vascular smooth muscle cells.