Diadenosine Homodinucleotide Products of ADP-ribosyl Cyclases Behave as Modulators of the Purinergic Receptor P2X7

Abstract

ADP-ribosyl cyclases from both vertebrates and invertebrates were previously shown to produce two isomers of P1,P2 diadenosine 5′,5′"-P1, P2-diphosphate, P18 and P24, from cyclic ADP-ribose (cADPR) and adenine. P18 and P24 are characterized by an unusual N-glycosidic linkage in one of the adenylic mononucleotides (Basile, G., Taglialatela-Scafati, O., Damonte, G., Armirotti, A., Bruzzone, S., Guida, L., Franco, L., Usai, C., Fattorusso, E., De Flora, A., and Zocchi, E. (2005) Proc. Natl. Acad. Sci. U.S.A. 102, 14509–14514). P24, but not P18, proved to increase the intracellular Ca²⁺ concentration ([Ca²⁺]_i) in HeLa cells and to negatively affect mitochondrial function. Here we show that micromolar P24, but not P18, triggers a slow and sustained influx of extracellular Ca²⁺ through the opening of the purinergic receptor/channel P2X7. On the other hand, P18 inhibits the Ca²⁺ influx induced by 0.6 mm ATP in HEK293 cells stably transfected with P2X7, with an IC₅₀ of ∼1 μm. Thus, P18 is devoid of intrinsic P2X7 stimulatory activity and behaves as an ATP antagonist. A P2X7-mediated increase of the basal [Ca²⁺]_i has been demonstrated to negatively affect Schwann cell (SC) function in rats with the inherited, peripheral neuropathy Charcot-Marie-Tooth 1A (CMT1A) (Nobbio, L., Sturla, L., Fiorese, F., Usai, C., Basile, G., Moreschi, I., Benvenuto, F., Zocchi, E., De Flora, A., Schenone, A., and Bruzzone S. (2009) J. Biol. Chem. 284, 23146–23158). Preincubation of CMT1A SC with 200 nm P18 restored the basal [Ca²⁺]_i to values similar to those recorded in wild-type SC. These results identify P18 as a new P2X7 antagonist, potentially useful in the treatment of CMT1A.

Introduction

CD38, a member of the ADP-ribosyl cyclase (ADPRC)³ family, is a multifunctional enzyme ubiquitously expressed in eukaryotic cells (1). ADPRCs metabolize NAD⁺ and produce several molecules that have been demonstrated to be involved in intracellular calcium signaling and in the regulation of Ca²⁺-dependent cell functions (1, 2). NAD⁺ is converted by ADPRCs to cyclic ADP-ribose (cADPR), a second messenger mobilizing Ca²⁺ from intracellular stores via ryanodine receptors (RyR) (1, 3). In addition to cADPR, ADPRCs generate the following products: (i) ADP-ribose (ADPR), which opens TRPM2 channels allowing extracellular Ca²⁺ influx in mammalian cells (4, 5); (ii) the ADP-ribose dimer (ADPR₂), which synergizes with cADPR in releasing Ca²⁺ from sea urchin egg microsomes (6); and (iii) nicotinic acid adenine dinucleotide 2′-phosphate (NAADP⁺), which induces calcium mobilization from intracellular stores, identified as acidic organelles or as endoplasmic reticulum in different mammalian cell types (7–9).

Recently, ADPRCs from Axinella polypoides (Porifera, Demospongiae), from Aplysia californica (Mollusc), and human CD38 have been demonstrated to synthesize three diadenosine homodinucleotides from cADPR and adenine diadenosine Ap2A and two isomers thereof (designated P18 and P24 from their HPLC retention times), which are characterized by an unusual N-glycosidic linkage between one of the adenines and ribose: C1′-N1 in P18 and C1′-N3 in P24 (10).

The presence of Ap2A has been reported in human platelets and in cardiac myocytes (11, 12). P18 and P24 were also identified in resting platelets and were shown to be released together with Ap2A following addition of thrombin (13). Ap2A, P18, and P24 inhibit platelet aggregation induced by thrombin, ADP, or collagen, through the activation of the purinergic receptor P2Y₁₁, resulting in an increase of the intraplatelet cyclic AMP (cAMP) concentration and subsequent overproduction of nitric oxide, the most potent endogenous antagonist of aggregation (13).

In mammalian cells, addition of extracellular Ap2A, P18 or P24 differently affects the [Ca²⁺]_i: whereas P18 induces its decrease, P24 conversely elicits an increase of the [Ca²⁺]_i, through extracellular Ca²⁺ influx, and Ap2A synergizes with cADPR in releasing Ca²⁺ from RyR (10). P24 showed a potent cytotoxic effect on several human cell lines via a mechanism involving mitochondrial dysfunction and the induction of apoptosis (10, 14). Micromolar concentrations of the dinucleotide induce, (i) dissipation of the mitochondrial proton gradient (ΔΨ_m) in intact cells, through opening of the permeability transition pore, and, (ii) inhibition of Complex I of the respiratory chain in isolated mitochondria (14). Dissipation of ΔΨ_m is a consequence of the mitochondrial Ca²⁺ increase, which follows the elevation of cytoplasmic Ca²⁺ increase caused by influx of extracellular Ca²⁺ (14).

Activation of the purinergic receptor P2X7, a ligand-gated channel, is known to induce cytotoxicity in mammalian cells, as a consequence of a massive Ca²⁺ influx through the opening of a large, nonselective membrane pore (15–17). In addition, prolonged stimulation of P2X7 by ATP leads to apoptosis, through influx of extracellular Ca²⁺, subsequent Ca²⁺ overload of the mitochondria and collapse of the ΔΨ_m (18, 19).

Here we identify activation of the purinergic receptor/channel P2X7 as the mechanism mediating the P24-induced Ca²⁺ influx in P2X7-expressing HEK293 cells. On the contrary, P18 did not induce Ca²⁺ entry through P2X7 but behaved as a potent ATP antagonist on P2X7. Interestingly, P18 restored normal [Ca²⁺]_i values in Schwann cells (SC) from rats with the Charcot-Marie-Tooth 1A (CMT1A) neuropathy, a demyelinating disease characterized by an increased basal [Ca²⁺]_i in SC, because of overexpression of P2X7 (20).

EXPERIMENTAL PROCEDURES

Materials

FLUO-3 AM and FURA-2 AM were obtained from Calbiochem (Milan, Italy). P18 and P24 were produced and purified as described in Ref. 10. Ap2A was synthesized from AMP and HPLC purified as described in Ref. 14. All other chemicals were obtained from Sigma.

Cell Culture

HEK293 cells stably transfected with human P2X7 (HEK293-P2X7 cells) and control HEK293 cells transfected with the empty vector (HEK293-mock cells), obtained as described in Ref. 18, and HeLa cells were cultured as were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with fetal calf serum (10%), penicillin (50 units/ml), and streptomycin (50 μg/ml) in a humidified 5% CO₂ atmosphere at 37 °C.

Primary SC cultures were isolated from sciatic nerves of adult CMT1A rats, as previously described (20). Control SC cultures were obtained from wild-type rats with the corresponding genetic background. SC from adult animals were grown for 4 days in DMEM/F12 (Invitrogen, Srl, Milano, Italy) containing 10% fetal calf serum, penicillin, streptomycin, and 10–5 m cytosine arabinoside (Ara-C) (Sigma).

Ca²⁺ Measurements

Adherent HeLa cells (2 × 10⁴ cells/well in 96-well plates) were loaded with 10 μm Fura-2AM for 45 min at 37 °C in complete medium. Cells were then washed with HBSS and Ca²⁺ measurements were performed in the same buffer. HEK293-P2X7 and HEK293-mock cells were incubated with 10 μm Fluo-3AM for 30 min at 25 °C in HBSS. Cells were then washed once with HBSS, resuspended in the same solution at 1 × 10⁶ cells/ml and seeded in 96-well plates (3 × 10⁴ cells/well). Fluo-3-loaded cells were then washed twice in Hank's Balanced salt solution (HBSS), and cell fluorescence was monitored at 1-min intervals on a Fluostar Optima microplate reader (BMG Labtechnologies, Offenburg, Germany: excitation, 485 nm; emission, 520 nm). For HEK293 cells, fluorescence intensity was monitored under stirring: before each reading cycle, an orbital shaking was made (1 s, 1 mm width, 350 rounds/min).

Adherent (on 20-mm poly-lysine-coated coverslips) HEK293 cells and wild type and CMT1A SC were incubated with 10 μm Fura-2AM in complete medium for 45 min at 37 °C and then washed with Hank's Balanced Salt Solution (HBSS). The coverslips were mounted on the stage of an inverted microscope (Zeiss IM35, Germany) and [Ca²⁺]_i calibrations were performed as described previously (21).

Mitochondrial Membrane Potential in Intact Cells

For measurements of mitochondrial membrane potential, cells were incubated in complete medium with 30 nm tetramethylrhodamine methyl ester (TMRM) for 2.5 h, in the presence or absence of 100 μm P24 or 3 mm ATP. Cells were then washed, resuspended in HBSS, and seeded on eight-well chambered cover glasses, precoated with poly-lysine, at a cell density of 1 × 10⁶ cells/well. After acquisition of the basal fluorescence level, the uncoupler carbonyl cyanide 4-trifluoro-methoxyphenylhydrazone (FCCP, 4 μm) was added. The difference between the basal fluorescence level and the fluorescence reached upon addition of FCCP was considered as a measure of ΔΨ_m (14). Cell fluorescence images were acquired on a Leica TCS SL confocal microscope, equipped with a HCX PL APO CS 63.0 × 1.40 oil objective, at 40-s intervals.

Measurement of Ethidium Bromide Entry

HEK293-P2X7 and HEK293-mock cells were seeded in 96-well plates (6 × 10⁴ cells/well) and incubated in the presence of 20 μm ethidium bromide. Ethidium bromide influx was evaluated with a fluorescence plate reader (Fluostar Optima; excitation and emission wavelengths were 355 and 390 nm, respectively), upon addition of different concentrations of either ATP, or P24, or P18 or Ap2A. To evaluate P24- or P18-mediated inhibition of ATP-induced uptake of ethidium bromide, cells were preincubated for 10 min with 20 μm P24 or P18, before addition of ATP. Fluorescence intensity was analyzed under stirring. Before each reading cycle, an orbital shaking was made (1 s, 1 mm width, 350 rounds/min).

Cytotoxicity Assays

HEK293-P2X7 and HEK293-mock cells (2 × 10⁴ cells/well), wild type and CMT1A SC (1.2 × 10⁴ cells/well) were seeded in quadruplicate in 96-well plates, and incubated (or not) with different concentrations of P18, P24, or Ap2A.

After 48 h of culture, cell proliferation was evaluated with the CellTiter 96® Aqueous One solution cell proliferation assay, which uses a solution containing a tetrazolium compound (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt; MTS) and an electron coupling reagent (phenazine ethosulfate) (Promega). Measurements of formazan absorbance at 450 nm were performed with a FluoStar Optima microplate reader. A standard curve of freshly seeded cells was prepared, stained in parallel and used to estimate the number of viable cells. Images of cells were taken with a Leica DFC290 digital camera. Lethal concentration 50% (LC₅₀) was calculated from the equation of the non-linear regression curves.

Statistical Analyses

All parameters were tested by paired t test or one-way ANOVA followed by Tukey test or Dunnett's method, as appropriate; p values <0.05 were considered significant.

RESULTS

P24 Activates the P2X7 Receptor

P24 has been demonstrated to trigger an increase of the [Ca²⁺]_i in HeLa cells, which was prevented by extracellular EDTA (10, 14). The P24-induced Ca²⁺ increase in these cells was also abrogated by addition of 10 μm Gd³⁺ (Fig. 1) or 30 μm La³⁺ (not shown), confirming that it was due to influx of extracellular Ca²⁺. Moreover, P24 induces apoptosis in HeLa cells because of dissipation of the ΔΨ_m, caused by mitochondrial Ca²⁺ overload (14).

FIGURE 1.

Effect of Gd³⁺ and of KN-62 on the P24-induced [Ca²⁺]_i increase in HeLa cells. P24 (20 μm, black traces) or the same volume of buffer (gray traces) were added: A, to untreated HeLa cells, or B, to cells preincubated with 10 μm GdCl₃ for 10 min, or C, to cells preincubated with 30 μm KN-62 for 30 min. The [Ca²⁺]_i increase in FURA-2-loaded HeLa cells was continuously recorded with a dual wavelength microfluorimetric detector allowing calibration of the [Ca²⁺]_i (21). Representative traces are shown (n = 4). D, Δ[Ca²⁺]_i were calculated for each condition as the difference between values at time = 900 s and time 0, and are shown as mean ± S.D. *, p < 0.001 by Dunnet's method, compared with the P24 bar.

The similar effects observed on cell Ca²⁺, ΔΨ_m, and viability upon activation of P2X7 (see Introduction) or administration of P24 (14), suggested to investigate whether P24 induced extracellular Ca²⁺ influx in HeLa cells through the opening of the P2X7 channel. Thus, HeLa cells were preincubated with KN-62, an inhibitor of P2X7, as well as of Ca²⁺-calmodulin-dependent protein kinase II (22), and then treated with P24. As shown in Fig. 1, KN-62 abrogated the P24-induced [Ca²⁺]_i increase, thus supporting the involvement of P2X7 in P24-stimulated Ca²⁺ influx.

To confirm a role for P2X7 in the P24-induced Ca²⁺ influx, HEK293 cells stably transfected with P2X7 (HEK293-P2X7 cells) were treated in parallel with either ATP or P24 at two different concentrations (100 μm and 1 mm). P24 induced a slow and sustained [Ca²⁺]_i increase already at 100 μm, a concentration at which ATP was ineffective (Fig. 2A), in agreement with previous data on the affinity of P2X7 for this dinucleotide (23). At the concentration of 1 mm, the [Ca²⁺]_i rise triggered by P24 was larger than at 100 μm but substantially smaller than that caused by ATP (Fig. 2A). A slow and sustained [Ca²⁺]_i increase induced by 100 μm P24 was also observed on single cells (not shown), suggesting that continuous presence of P24 does not induce P2X7 desensitization, in line with what observed with ATP (17).

FIGURE 2.

P24- and ATP-induced [Ca²⁺]_i increase in HEK293-P2X7 cells. A, P24 (0.1 mm, white squares, or 1 mm, black squares), or ATP (0.1 mm, white triangles, or 1 mm, black triangles) was added to Fluo 3-loaded HEK293-P2X7 cells. Fluorescence changes were measured in parallel using a fluorescence plate reader. Addition of up to 1 mm Ap2A or P18 did not induce any Ca²⁺ increase (not shown). Inset, the variations of fluorescence, calculated for each condition as the difference between the maximum value and time 0, are shown as mean ± S.D. *, p < 0.0001 by Tukey test, compared with control (c, addition of buffer); #, p < 0.0001 by Tukey test, compared with 1 mm P24. B, [Ca²⁺]_i increase in Fura-2-loaded HEK293-P2X7 cells following addition of the indicated concentrations of P24 was recorded as in (21). Inset, the Δ[Ca²⁺]_i, calculated for each condition as the difference between values at time 10 min and time 0, are shown as mean ± S.D. *, p < 0.001 by Tukey test, compared with control (c, addition of buffer). #, p < 0.01 by Tukey test, compared with 5 μm P24, §, p < 0.01 by Tukey test, compared with 20 μm P24. C, HEK293-P2X7 cells were preincubated (black circles), or not (black squares) with 20 μm KN-62 for 10 min prior to addition of 100 μm P24. White squares indicate the addition of 100 μm P24 to HEK293-mock cells. Fluorescence changes were measured in parallel using a fluorescence plate reader. Inset, the variations of fluorescence, calculated for each condition as the difference between values at time 600 s and time 0, are shown as mean ± S.D. #, p < 0.001 by t test, compared with the P24 bar (untreated HEK293-P2X7 cells, addition of 100 μm P24). Representative traces are shown (A, n = 8; B, n = 4; C, n = 3).

The effect of P24 was concentration-dependent, with as low as 5 μm being already effective, as measured at 10 min after the addition (Fig. 2B). Preincubation of HEK293-P2X7 cells with KN-62 prevented the P24-induced [Ca²⁺]_i increase and P24 had no effect on [Ca²⁺]_i homeostasis in HEK293 cells transfected with the empty vector (HEK293-mock cells) (Fig. 2C). Neither of the two P24 isomers, P18 or Ap2A, elicited a [Ca²⁺]_i increase in HEK293-P2X7 cells or in HEK293-mock cells (not shown).

To investigate the involvement of P2X7 in P24-stimulated ΔΨ_m collapse, HEK293-P2X7 cells were loaded for 2.5 h with the mitochondrial fluorescent probe TMRM, in the presence or absence of 100 μm P24. The difference between the basal fluorescence level and the fluorescence recorded upon addition of the uncoupler FCCP revealed an almost complete (85%) dissipation of the ΔΨ_m (Fig. 3A). The ΔΨ_m of cells incubated with 1 mm ATP was also dissipated to a similar extent (by 75%, Fig. 3A), in agreement with previous observations (18).

FIGURE 3.

Dissipation of the ΔΨ_m and influx of ethidium bromide into HEK293-P2X7 cells. A, HEK293-P2X7 cells were loaded for 2.5 h with 30 nm TMRM, in the presence or absence of 100 μm P24 or 1 mm ATP. Data are the difference between the fluorescence level recorded before and after addition of the uncoupler FCCP, and are the mean ± S.D. of four experiments. *, p < 0.005 by Tukey test, compared with control. B, ethidium bromide influx into HEK293-P2X7 cells was monitored with a fluorescence plate reader, as described under “Experimental Procedures,” in the absence of extracellular nucleotides (white squares), or upon addition of 1 mm Ap2A (black circles), P24 (black triangles), or ATP (black squares). Representative traces are shown (n = 3). Inset, the variations of fluorescence, calculated for each condition as the difference between values at time 600 s and time 0, are shown as mean ± S.D. *, p < 0.001by Tukey test, compared with control.

Activation of P2X7 initially leads to increased membrane permeability to small cations, followed by the progressive opening of a large, nonselective membrane pore, permeable to large fluorescent dyes (15–17). Addition of up to 1 mm P24 to HEK293-P2X7 cells did not cause influx of either Lucifer yellow or of ethidium bromide. Conversely, 1 mm ATP induced cell uptake of Lucifer yellow (not shown) and of ethidium bromide (Fig. 3B). Altogether, these results indicate that the P24-induced activation of P2X7 causes opening of the ion channel, allowing subsequent Ca²⁺ influx, but is not followed by pore formation.

Influx of extracellular Ca²⁺, and the subsequent dissipation of ΔΨ_m, was previously shown to be responsible for P24-induced cell death (14). Thus, we explored P24 cytotoxicity in HEK293-P2X7 and in HEK293-mock cells. Up to 10 μm P24 had no effect on viability of HEK293-mock cells, whereas it showed a concentration-dependent cytotoxicity on HEK293-P2X7 cells (Fig. 4, A and B), with an LC₅₀ of 1.1 ± 0.2 μm, ∼10-fold lower than that reported for HeLa cells (9.8 μm, Ref. 10). After 24 h of incubation in the presence of 0.2, 1, or 5 μm P24, [Ca²⁺]_i levels in HEK293-P2X7 cells were concentration dependently increased, from a basal value of 43 nm, to 83, 102, and 131 nm, respectively (inset to Fig. 4A). In HEK293-P2X7 cells preincubated in the presence of 100 nm KN-62, a concentration inhibiting P2X7, without affecting CaMKII (22), the P24-induced cytotoxicity was almost completely prevented (Fig. 4C), suggesting that P24-induced cell death is mediated by P2X7 activation. To confirm the involvement of P2X7 in the P24-induced cytotoxicity, HEK293-P2X7 and -mock cells were preincubated in the presence of a different inhibitor of this purinergic receptor, oxidized ATP (oATP) (17): as shown in Fig. 4C, oATP almost completely prevented P24-induced HEK293-P2X7 cell death.

FIGURE 4.

P18 and P24 cytotoxicity on HEK293-P2X7 and -mock cells. A, HEK293-P2X7 cells (black squares) or HEK293-mock cells (white squares) were cultured for 48 h in the absence (control) or in the presence of different concentrations of P24. Cell viability was evaluated with the MTS colorimetric method. Results are expressed as percentage of live cells relative to control; mean ± S.D. from 11 different experiments are shown. #, p < 0.001 by Tukey test, compared with untreated HEK293-P2X7 cells. Inset: HEK293-P2X7 cells were cultured for 24 h in the presence of the indicated concentrations of P24 and then loaded with Fura-2AM. The [Ca²⁺]_i was evaluated as described (21). *, p < 0.005 by Tukey test, compared with control (in the absence of P24); #, p < 0.05 by Tukey test, compared with 0.2 and 1 μm P24 (n = 3). B, HEK293-P2X7 cells or HEK293-mock cells were cultured for 48 h in the absence (control) or in the presence of the indicated concentrations of P24 and phase-contrast light microscopy images were acquired. One representative experiment is shown. C, HEK293-P2X7 cells (black bars), or HEK293-mock cells (white bars) were preincubated (or not) for 1 h with 100 nm KN-62, or with 100 μm oxidized ATP (oATP), and then cultured for 48 h in the absence (control) or in the presence of 1 μm P24. Cell viability was evaluated with the MTS method. Results are expressed as percentage of live cells relative to control; the mean ± S.D. from six different experiments is shown. *, p < 0.005 by Dunnett's method, compared with HEK293-P2X7 cells treated with P24. D, HEK293-P2X7 cells (black squares) or HEK293-mock cells (white squares) were cultured for 48 h in the absence (control) or in the presence of different concentrations of P18. Cell viability was evaluated with the MTS method. Results are expressed as percentage of live cells relative to control; mean ± S.D. from six different experiments are shown.

P18 was also previously found to decrease cell viability on several human cell types, although at a concentration 1-log higher than P24, via an as yet unidentified mechanism (10). P18-induced cytotoxicity was not significantly different (p > 0.5) on HEK293-P2X7 and on HEK293-mock cells, with an LC₅₀ of 25 ± 5 and 30 ± 4 μm, respectively (Fig. 4D), demonstrating that it does not depend on expression of P2X7. On the other hand, Ap2A did not significantly affect viability of either HEK293-P2X7 or HEK293-mock cells at concentrations up to 100 μm (not shown).

P24 and P18 Antagonize the Opening of P2X7 by ATP

The observation that P24 triggered a Ca²⁺ rise in the absence of pore formation (Figs. 2 and 3), suggested that it might be an incomplete P2X7 agonist, and prompted experiments aimed at investigating its effects on ATP-dependent P2X7 activation. To this purpose, HEK293-P2X7 cells were preincubated for 10 min with P24 and then challenged with increasing concentrations of ATP. In the presence of 100 μm P24, the Ca²⁺ increase induced by ATP was significantly reduced by 100, 66, 58, or 13% at the ATP concentrations of 0.3, 0.45, 0.6, or 1 mm, respectively (Fig. 5A). The Ca²⁺ rise induced by 0.6 mm ATP was concentration dependently inhibited by P24, with an IC₅₀ of 17.0 ± 3.4 μm, as calculated from the logarithmic regression curves (Fig. 5C). Although 20 μm P24 per se induced a slow Ca²⁺ increase, it reduced the ATP-induced Ca²⁺ rise by ∼50% (inset to Fig. 5C, black trace).

FIGURE 5.

P24 and P18 antagonize the opening of P2X7 by ATP. A, Fura2-loaded HEK293-P2X7 cells were preincubated, or not (control gray bars), for 10 min with 100 μm P24 (black bars) or 2 μm P18 (white bars) prior to addition of the indicated concentrations of ATP. The maximal [Ca²⁺]_i reached in each condition upon addition of ATP is shown. Results are mean ± S.D. from at least four different determinations. #, p < 0.001; §, p < 0.0001 by t test, compared with the corresponding addition of ATP, in the absence of P24 or P18. B, data obtained as in panel A, expressed as percentage of maximal Ca²⁺ response (gray squares, ATP; black squares, P24; white squares, P18). In addition, effect of 10 min of preincubation with 20 μm P18 is shown (white circles). S.D.<25% (not shown) C, Fluo 3-loaded HEK293-P2X7 cells were preincubated for 10 min with the indicated concentrations of P24, prior to addition of 0.6 mm ATP. Inset, representative traces are shown: 20 μm P24 was added (black squares) or not (white squares) at time 0 to HEK293-P2X7 cells, that were stimulated after 10 min with ATP. D, Fluo 3-loaded HEK293-P2X7 cells were preincubated for 10 min with the indicated concentrations of P18, prior to addition of 0.6 mm ATP. Inset, representative traces are shown: 1 μm P18 was added (black squares) or not (white squares) at time zero to HEK293-P2X7 cells, that were stimulated after 10 min with ATP. Fluorescence changes were recorded with a fluorescence plate reader. In both B and C, results are expressed as % of inhibition of the fluorescence increase induced by the same ATP concentration, in the absence of dinucleotides. Means ± S.D. of at least three different experiments are shown. E, HEK293-P2X7 cells were preincubated (or not, white squares) with 20 μm P24 (black squares) or P18 (white circles) for 10 min, before addition of 1 mm ATP (arrow). Ethidium bromide influx into the cells was monitored with a fluorescence plate reader, as described under “Experimental Procedures.” Representative traces are shown (n = 3). Inset, the variation of fluorescence, calculated for each condition as the difference between values at time 900 s and time 0, is shown as mean ± S.D. *, p < 0.0001 by Dunnett's method, compared with ATP (addition of ATP, in the absence of P24 or P18).

Next, we explored whether P18, which by itself had no effect on [Ca²⁺]_i homeostasis, could affect the ATP-induced Ca²⁺ increase in HEK293-P2X7 cells. Indeed, preincubation of the cells for 10 min with 50 μm P18, completely abolished the Ca²⁺ influx induced by 1 mm ATP (not shown). Fig. 5A shows the percentages of inhibition obtained upon preincubation of HEK293-P2X7 cells with 2 μm P18, in comparison with those obtained with 100 μm P24. The IC₅₀ of P18 on the [Ca²⁺]_i rise induced by 0.6 mm ATP was 1.2 ± 0.2 μm, as calculated from the logarithmic regression curves (Fig. 5D). Thus, P18 inhibits the effect of ATP on P2X7, much more efficiently than P24, without affecting per se the [Ca²⁺]_i (inset to Fig. 5D, black trace).

The concentration-effect curves for ATP were progressively shifted to the right in the presence of 100 μm P24 or of 2 and 20 μm P18 (Fig. 5B), without decreasing agonist efficacy. These data indicate that both P24 and P18 blocks ATP-evoked activation of P2X7 receptors in a competitive fashion. A 10-min preincubation of the cells in the presence of 100 μm Ap2A did not significantly decrease the [Ca²⁺]_i rise induced by 1 mm ATP (not shown).

Next, we investigated whether P24 and P18 could affect the delayed pore formation following exposure to extracellular ATP. Fig. 5E shows that addition of either P24 or P18 (20 μm) completely abrogated the opening of the pore induced by 1 mm ATP in HEK293-P2X7, as measured by ethidium bromide uptake.

P18 Restores Normal Basal [Ca²⁺]_i in CMT1A SC

A P2X7-mediated increase of the basal [Ca²⁺]_i has been demonstrated to cause functional derangement in SC from rats with the CMT1A neuropathy and pharmacological P2X7 antagonists restored normal [Ca²⁺]_i in CMT1A SC (20). To investigate the effect on CMT1A SC of the newly identified P2X7 antagonist P18 on CMT1A SC, we first evaluated its toxicity on CMT1A SC. As shown in Fig. 6A, P18 was devoid of cell toxicity at concentrations ≤25 μm, Ap2A did not affect cell viability at concentrations up to 100 μm, whereas the LC₅₀ of P24 was 22.3 ± 3.2 μm after 48 h of incubation, as calculated from the non-linear regression curve. Similar results for P24, P18, and Ap2A were obtained on SC from wild-type rats (not shown) and the LC₅₀ values for P18 and P24 on SC (∼118.0 and 22.3 μm, respectively) are similar to those already reported in HeLa cells (10). BzATP was chosen to stimulate P2X7 in SC, since it is the most potent agonist for this purinergic receptor, with a much higher affinity than ATP (24). P18 antagonized the BzATP-induced [Ca²⁺]_i rise in CMT1A SC in a concentration-dependent manner (Fig. 6B), with an IC₅₀ of ∼0.1 μm, as determined from the logarithmic regression curve. As shown in Fig. 6C, an 18-h preincubation of CMT1A SC in the presence of 0.05, 0.2, or 1 μm P18 reduced the basal [Ca²⁺]_i of CMT1A SC to values (77, 60, and 44 nm, respectively) progressively closer to those measured in wild-type SC (53 ± 7 nm). Conversely, preincubation for 18 h with 5 μm Ap2A did not modify the [Ca²⁺]_i of CMT1A (Fig. 6C).

FIGURE 6.

P18 restores normal basal [Ca²⁺]_i in CMT1A SC. A, CMT1A SC were cultured for 48 h in the absence (control) or in the presence of different concentrations of P18 (white bars), Ap2A (gray bars), or P24 (black bars). Cell viability was evaluated with the MTS colorimetric method. Results are expressed as percentage of live cells relative to control; the mean ± S.D. of three experiments is shown. *, p < 0.05; **, p < 0.005 by Tukey test, compared with control. B, Fluo 3-loaded CMT1A SC were preincubated for 30 min with the indicated concentrations of P18, prior to addition of 100 μm BzATP. Fluorescence changes were recorded with a fluorescence plate reader. Results are expressed as percentage of inhibition of the maximal fluorescence increase induced by the same BzATP concentration, in the absence of the dinucleotide. The mean ± S.D. of four different experiments is shown. C, CMT1A SC were cultured for 18 h in the absence (control) or in the presence of the indicated concentrations of P18, or of 5 μm Ap2A, and then loaded with Fura-2AM. The [Ca²⁺]_i was evaluated as described (21). The mean ± S.D. of three different experiments is shown. *, p < 0.05; **, p < 0.005 by Dunnett's method, compared with control.

DISCUSSION

In this study, we identify P18 and P24, two adenylic dinucleotides present in mammalian cells (10), as modulators of the purinergic receptor/channel P2X7. Both ADPRC products are active at much lower concentrations than those required for ATP, the only endogenous nucleotide activator of this receptor/channel known so far (23). However, conformational changes produced by pH, ionic or membrane lipid composition have been reported to induce allosteric modulation of P2X7 affinity for ATP (25–28). Recently, other physiological ligands have been reported to activate P2X7 at lower concentrations than ATP. Extracellular NAD⁺ has been reported to regulate P2X7 opening through two different mechanisms in murine cells: in T lymphocytes, NAD⁺ induces the ATP-independent activation of P2X7, via ADP-ribosylation of arginine residues on the P2X7 ectodomain (29); in macrophages, the NAD⁺-dependent ADP-ribosylation of P2X7 synergizes with ATP to open P2X7 (30). The human antimicrobial peptide cathelicidin LL-37 has been suggested to represent a possible non-nucleotide physiological ligand for P2X7: LL-37 induces caspase-1 activation and secretion of IL-1β in LPS-primed monocytes, inhibits apoptosis in human neutrophils and stimulates fibroblast proliferation, through the activation of P2X7 (31–33).

Activation of P2X7 by P24 causes influx of extracellular Ca²⁺ (Fig. 2), ΔΨ_m dissipation (Fig. 3A) and cell death (Fig. 4) in HEK293 cells transfected with P2X7, at nucleotide concentrations inactive on mock-transfected HEK293. The events occurring upon stimulation of HEK293-P2X7 cells with P24 have been described to take place also in HeLa cells (14). The fact that KN-62 abrogated the P24-induced Ca²⁺ influx (Fig. 1), suggested the involvement of P2X7. Besides expressing the full-length P2X7, HeLa cells also express a naturally occurring truncated variant, named P2X7-j (34). Importantly, in HEK293cells overexpressing both P2X7-j and P2X7, the P2X7-j was shown to interact with P2X7 receptor, leading to inhibition of the Bz-ATP-induced Ca²⁺ influx and pore formation (34). Thus, the effect of P24 on the Ca²⁺ influx through P2X7 in HeLa cells could be underestimated because of the interaction between endogenous P2X7 and P2X7-j, although the influence of the latter on the effect of P24 on P2X7 in cells not overexpressing the truncated form remains to be established. In addition to activating P2X7 in HEK293 cells, P24 also decreases the Ca²⁺ influx triggered by 0.3–0.6 mm ATP (Fig. 5, A–C) and prevents the pore formation induced by ATP (Fig. 5E).

P2X7 is characterized by two states of permeability (16, 35): at high micromolar concentrations of ATP, P2X7 behaves as a cation-selective channel; a prolonged exposure to ATP triggers the transition to a non-selective pore, permeable to molecules of mass up to 900 Da (15, 36). The molecular mechanisms leading to P2X7-pore formation are still unclear: pore formation could be an intrinsic property of P2X7, through its oligomerization, or it could be due to recruitment by activated P2X7 of other channels (17, 37). However, regardless of the molecular composition of the pore, to the best of our knowledge, P24 represents the first molecule endowed with the ability to open the cation channel without causing pore formation (Figs. 2 and 3B). The molecular mechanism underlying this peculiar feature of P24 has not been investigated in the present study. P24 may differently affect P2X7 oligomer formation or stabilization. A similar explanation has been suggested for the differential effect of polymyxin B (PMB) and polymyxin B nonapeptide (PMBN) on P2X7 permeability. In contrast to PMB, PMBN is unable to potentiate the ATP-induced Ca²⁺ increase and pore formation, possibly because PMBN is unable to enhance P2X7 oligomerization (38).

The fact that P24 triggers opening of the P2X7 cation channel (Fig. 2), while not leading to pore formation (Fig. 3B), indicates that the P2X7-induced cell death is not necessarily a consequence of non-selective permeability to molecules upon pore formation. Instead, the P24-induced loss of mitochondrial proton gradient (Fig. 3A) and cell death (Fig. 4) appear to be a consequence of extracellular calcium influx through the cation channel. This conclusion is in line with the observation that the addition of ATP causes a collapse of ΔΨ_m in HEK293 cells transfected with a truncated form of the P2X7 receptor, which retains channel activity but cannot form the non-selective pore (18). Therefore, P24 could represent a useful molecular tool to identify the molecular events downstream of P2X7 activation that require pore formation.

P18, an isomer of P24, does not activate P2X7, as indicated by the fact that it does not induce any Ca²⁺ increase in HEK293-P2X7 cells (see legend to Fig. 2A), it does not affect mitochondrial function (14) and it is equally toxic on HEK293-P2X7 and on HEK293-mock cells (Fig. 4D), this demonstrating that P18-induced cytotoxicity does not depend on expression of P2X7. The mechanism underlying P18-induced cytotoxicity in HEK293 cells, as well as in other cell types (10), remains to be identified.

Notably, P18 antagonizes the ATP-induced Ca²⁺ rise and pore formation (Fig. 5), thus representing a new, endogenous, negative modulator of P2X7. The IC₅₀ values calculated for P18 as inhibitor of the ATP-induced Ca²⁺ rise in human and in rat cells are quite different (1 and 0.1 μm, respectively, in HEK293-P2X7 and in CMT1A SC, Figs. 5D and 6B). This discrepancy may be due to the fact that a different P2X7 agonist was used (ATP versus BzATP) or to the already reported different selectivity of human and rat cells to several P2X7 antagonists (39).

Increasing effort is devoted to the search of antagonists of the P2X7 receptor, due to its role in regulating several key inflammatory molecules, including IL-1β, IL-18, TNF-α, and IL-6 (40), and to its role in pain perception. Thus, P2X7 antagonists are likely to have therapeutic potential as novel anti-inflammatory and anti-nociceptive drugs (reviewed in Refs. 40, 41). Indeed, local administration of the P2X7 antagonist oxidized-ATP reduced inflammation-induced mechanical hyperalgesia in rats (42, 43). Moreover, systemic administration of P2X7-selective antagonists produced antinociceptive effects in models of neuropathic and inflammatory pain (44–46). New studies on P2X7 knock-out mice indicate that this receptor plays an important role in a number of other chronic inflammatory conditions, including multiple sclerosis and hepatitis (47, 48).

Derangement of the [Ca²⁺]_i in CMT1A SC, which is caused by the inappropriate, chronic activation of the P2X7 channel by overexpression of the membrane protein PMP22 (20), is known to be responsible for the impairment of several important functional activities of SC, such as migration, production of ciliary neurotrophic factor (CNTF) and myelination (20). Results presented here indicate that P18 is capable of restoring normal basal values of the [Ca²⁺]_i in CMT1A SC, at concentrations (50 nm) devoid of cytotoxicity (Fig. 6, A and C), mandating further exploration of the effect of this dinucleotide on CMT1A SC functional activities.

The therapeutic value of modulators of the P2X7 receptor is not restricted to its antagonists: indeed, new P2X7 agonists may prove useful as anticancer agents, inducing tumor cell death (40). In this respect, P24 may represent a promising pharmacological tool, because of its high cytotoxicity, particularly on hemopoietic cells (10). The fact that P24 also prevents the P2X7-mediated, ATP-induced Ca²⁺ rise and pore formation (Fig. 5), renders P24 a double-faced molecule, active both as an agonist and as an ATP antagonist of P2X7. This might prove advantageous in anticancer therapy, due to the established role of chronic inflammation as a mechanism favoring cancer cell development and/or survival (40).

Despite their structural similarity, the three isomers P24, P18 and Ap2A exert very different effects on the P2X7 receptor: P24, but not P18 or Ap2A, induces the cation channel opening. Both P18 and P24 antagonize the ATP-induced Ca²⁺ increase, whereas Ap2A does not. The sigmoidal dose-response curves obtained with ATP, in the absence or presence of P18 or P24 (Fig. 5B), indicate that both P18 and P24 behave as competitive antagonists of ATP.

Admittedly, selectivity of these dinucleotides on the other members of the purinergic receptor family requires further investigation, although it is interesting to underline that P24, P18, and Ap2A, are all agonists of the P2Y₁₁ purinergic receptor (13).

All the three adenylic dinucleotide isomers are generated in mammalian cells by the ectocellular ADPRC CD38, in the presence of NAD⁺ and adenine (10, 49). This fact raises the intriguing possibility that availability of extracellular NAD⁺ and adenine may affect the activation state of P2X7 by enabling the ectocellular synthesis of its modulators P18 and P24. Indeed, both dinucleotides are present at micromolar concentrations (∼50 μm) in the supernatant of activated platelets, suspended at a concentration similar to that present in whole blood (13). The intracellular concentration of P24 and P18 is in the micromolar range in platelets (13) and CD38⁺ HeLa cells (10). These data suggest that concentrations of P24 and P18 capable of affecting the activation state of P2X7, could be reached “in vivo.”