P2X7 Receptor-Dependent and -Independent T Cell Death Is Induced by Nicotinamide Adenine Dinucleotide
http://www.100md.com
免疫学杂志 2005年第4期
Abstract
Adding NAD to murine T lymphocytes inhibits their functions and induces annexin V binding. This report shows that NAD induces cell death in a subset of T cells within seconds whereas others do not die until many hours later. Low NAD concentrations (<10 μM) suffice to trigger rapid cell death, which is associated with annexin V binding and membrane pore formation, is not blocked by the caspase inhibitor Z-VADfmk, and requires functional P2X7 receptors. The slower induction of death requires higher NAD concentrations (>100 μM), is blocked by caspase inhibitor Z-VADfmk, is associated with DNA fragmentation, and does not require P2X7 receptors. T cells degrade NAD to ADP-ribose (ADPR), and adding ADPR to T cells leads to slow but not rapid cell death. NAD but not ADPR provides the substrate for ADP-ribosyltransferase (ART-2)-mediated attachment of ADP-ribosyl groups to cell surface proteins; expression of ART-2 is required for NAD to trigger rapid but not slow cell death. These results support the hypothesis that cell surface ART-2 uses NAD but not ADPR to attach ADP-ribosyl groups to the cell surface, and that these groups act as ligands for P2X7 receptors that then induce rapid cell death. Adding either NAD or ADPR also triggers a different set of mechanisms, not requiring ART-2 or P2X7 receptors that more slowly induce cell death.
Introduction
The addition of NAD to murine T lymphocytes inhibits their functions (1) and induces annexin V binding (2, 3, 4). Treating naive T cells with NAD before injection into semiallogeneic mice blocks induction of graft-vs-host disease (2). Treating T cells before injection into syngeneic mice blocks their migration to the spleen and directs them to the liver where they appear to die (2). Injecting NAD directly into mice increases the number of apoptotic T cells in the liver (2). Taken together, these results indicate extracellular NAD (ecto-NAD) could play a role in immune regulation.
Several hypotheses have emerged to explain actions of ecto-NAD on T cells. One hypothesis is that it serves as a substrate for an arginine-specific cell surface ADP-ribosyltransferase (ART-2),3 which by ADP-ribosylating cell surface proteins causes the observed effects (1). Several observations support this hypothesis; anti-CD3 activation of murine T cells releases ART-2 and causes the cells to become resistant to NAD (5, 6), and using recombinant means, deletion of murine T cell ART-2 isoforms (ART2a and ART2b) inhibits cell surface ADP-ribosylation (7). Several of the proteins incorporating ADP-ribosyl groups have been identified, including LFA-1, CD27, CD43, CD45, and CD8 (8, 9), raising questions as to whether modification of one or more of these proteins induces cell death. It has been proposed that effects of ecto-NAD may arise from an activating ADP-ribosylation of purinergic receptors (10). It has also been proposed that ecto-NAD leads to an inactivating ADP-ribosylation of CD38, a cell surface NAD-glycohydrolase that can synthesize cyclic ADP-ribose (ADPR) (3). Yet another hypothesis arises from the observation that ecto-NAD inhibits mitogen-stimulated proliferation of rat T cells, which express little or no ART-2, and that, for rat T cells, adding ADPR is as effective as adding NAD. Because rat T cells degrade NAD to ADPR, the hypothesis arose that ADP-ribosylation may not be involved in the effects arising from the addition of NAD to cultured T cells but rather that NAD breakdown products act as ligands for purinergic receptors (11).
The data reported herein prompt a unifying model consistent with the above observations. The data demonstrate that adding NAD to T cells induces both rapid and slow death involving different sets of mechanisms. Rapid death involves pore formation and signaling through P2X7 receptors while slower death requires caspase activation but not P2X7 receptors. It is suggested that P2X7 engagement of ADP-ribosyl groups, attached by action of ART-2 to cell surface proteins, triggers receptor function and rapid cell death. This mechanism may provide a means by which NAD released during trauma controls T cell functions.
Materials and Methods
Mouse strains, T cell purification, and flow cytometric analysis
Pathogen-free female C57BL/6 (B6), and BALB/c mice, 6–8 wk of age were obtained from The Jackson Laboratory. NOD.129S4(B6)-Art2aTm1FknArt2bTm1Fkn (henceforth denoted as ART-2–/– mice) were produced at The Jackson Laboratory by outcross to a previously described B6 stock carrying targeted mutations in the tandem genes encoding the ART-2.1 and ART-2.2 ectoenzymes (7). Following nine backcrosses to NOD/Lt (N10) and verifying presence of NOD alleles at all known "Idd" loci, intercrosses were initiated, and ART2–/– homozygous breeding stock were received and bred at the University of Southern California animal facility (Los Angeles, CA). B6 P2X7–/– mice were kindly provided by Dr. C. Gabel (Ann Arbor, MI) and Pfizer and were bred at the University of Southern California animal facility (12). T cells were purified from spleen cells by nylon wool nonadherent (NWNA) cells and cultured in complete RPMI 1640 medium, containing 10% FBS (2). By FACS analysis, 85–90% of the NWNA cells were CD3 positive. For FACS analysis, T cells were preincubated with anti-mouse CD16/CD32 (2.4G2) mAb from BD Biosciences to block FcRs, and then incubated with various mAbs for 30 min at 4°C (2). Phenotypic analysis was performed using PE Cy5-conjugated anti-mouse CD3 (145-2C11) (BD Biosciences). To monitor induction of cell death, cells were stained with the Annexin V-FITC Apoptosis Detection Kit I (BD Biosciences). To demonstrate DNA fragmentation, cells were assayed using the TUNEL APO-Direct Apoptosis Assay kit (BD Biosciences). To quantitate cell surface ADP-ribosylation, NWNA cells (4 x 106 cells/ml in RPMI 1640) were incubated with etheno-NAD (Sigma-Aldrich), followed by incubation with etheno-ADPR-specific Ab 1G4, kindly provided by Dr. R. Santella (Mailma School of Public Health of Columbia University, New York, NY) (13) and a FITC goat anti-mouse Ig (BD Biosciences). FACS analysis was performed on a FACSCalibur (BD Biosciences).
Quantification of NAD metabolites
NWNA B6, BALB/c, and ART2–/– T cells (2 x 106) were incubated in 50 μl of serum-free PBS containing 200 μM NAD (Sigma-Aldrich) and 0.1 μCi [32P]NAD (Amersham) at 37°C. After centrifugation, supernatant from each sample (2 μl) was removed and applied to a 20 x 20 cm PEI-cellulose F glass-backed TLC plate (EMD Chemicals). TLC was developed in 0.3 M LiCl (Sigma-Aldrich) and exposed to a phosphor screen (Amersham) for 3 h at 25°C. NAD and ADPR were visualized using a PhosphorImager 445SI (Amersham), and radioactivity was quantified by ImageQuant 5.0 (Amersham) (11).
Western blot analyses
NWNA cells were incubated in RPMI 1640 with or without etheno-NAD (300 μM) for 30 min at 37°C. After washing twice in ice cold PBS, cells were re-suspended in lysis buffer (108 cells/ml PBS containing 1% Igepal, 1 mM AEBSF) and incubated for 30 min at 4°C. Insoluble material was removed by centrifugation (15 min, 10,000 x g), and supernatants were mixed with sample buffer, heated for 8 min at 95°C, and then loaded on SDS-PAGE gels. After electrophoresis, proteins were transferred onto PVDF membrane (Bio-Rad). Membranes were blocked with blocking buffer (LI-COR) for 1 h at room temperature. Blots were then incubated with 1G4 Ab (13) (ascites 1/500 dilution in TBS, 1% BSA, 0.05% Tween 20) for 2 h at room temperature. Membranes were washed with 0.05% Tween 20 in TBS and incubated with a 1/5000 dilution of Alexis Fluor 680-conjugated secondary anti-mouse IgG Abs (Molecular Probes) for 1 h. Protein bands were visualized by an infrared image system (LI-COR).
Cell proliferation and cell death assays and inhibition of caspases and P2X7 receptor function
To assay cell proliferation, NWNA cells were cultured on anti-CD3-coated tissue culture plates (2). Plates were incubated with a 1/1000 dilution of anti-CD3 Ab (500AA2 ascites) overnight. After washing the plates, T cells (5 x 105/well) were added in complete RPMI 1640 medium containing 10% FBS, and incubation was continued for 48 h. [3H]Thymidine (0.5 μCi/well) was added during the last 18 h of culture.
To assay induction of cell death, NWNA T cells were incubated with or without NAD, etheno-NAD, ADPR, or ATP (Sigma-Aldrich) for various times and assayed for propidium iodide (PI) uptake, annexin V, or TUNEL staining. To assay short-term effects of NAD (5 min), a 10-fold excess ice-cold PBS was added to the cell suspension to dilute NAD. Following centrifugation, cells were washed twice in ice-cold PBS before culture in complete RPMI 1640 medium. Assays up to 2 h were done in complete RPMI 1640 medium without FBS, while those for 24 h contained 10% FBS. Caspase inhibitor Z-VADfmk (BD Biosciences) was used at 50 μM and added to cell cultures 1 h before addition of NAD (14). To inhibit P2X7 receptors, 20 μM KN-62 (Sigma-Aldrich) dissolved in 0.01% DMSO were added 10 min before addition of NAD or ATP (15).
Results
NAD inhibits cell proliferation and induces a death signal in both BALB/c and B6 T cells
Adding NAD to either B6 or BALB/c T cells leads to cell surface ADP-ribosylation (2, 4) and increases annexin V staining, but there are significant differences. B6 T cells undergo more extensive cell surface ADP-ribosylation than BALB/c T cells (1, 4), but a higher proportion of BALB/c T cells is induced for cell death (4). These differences become evident when purified T cells are incubated with ART-2 substrate etheno-NAD and then stained with etheno-ADPR-specific Ab 1G4 (13) (Fig. 1A). BALB/c T cells show staining when compared with cells from ART-2 gene-deleted mice (7), but it is much less pronounced then that of B6 T cells. Concordant results are seen in immunoblots. Cell extracts from BALB/c show barely detectable bands, whereas those from B6 T cells show strong bands (Fig. 1B). These differences between B6 and BALB/c T cells provide an experimental system to study the pathways responsible for NAD-induced cell death.
FIGURE 1. ADP-ribosylation of T cells and effects of NAD, ADPR, and nicotinamide on T cell proliferation. A, ADP-ribosylation of B6, BALB/c, and ART-2–/– T cells. T cells were incubated with 300 μM etheno-NAD for 30 min, stained with Ab 1G4, and analyzed by FACS. B, Cell lysates were prepared from etheno-NAD treated cells separated by PAGE and analyzed by immunoblotting using 1G4 Ab. C, Inhibition of T cell proliferation by NAD, ADPR, and nicotinamide. B6 and BALB/c T cells were cultured on anti-CD3-coated tissue culture plates with the indicated concentrations of NAD, ADPR, and nicotinamide for 48 h. [3H]TdR was added for the last 18 h of incubation, and incorporation of 3H was determined. Error bars indicate SD. All data shown are from one of three repeat experiments.
Effects of NAD on cell proliferation were compared with T cells from B6 and BALB/c mice. Fig. 1C shows that anti-CD3 stimulated [3H]TdR incorporation into B6 and BALB/c T cells is almost completely inhibited by adding NAD at concentrations between 10 and 1000 μM. Therefore, effects of NAD on T cell proliferation are not proportional to the extent of cell surface ADP-ribosylation.
Next, the ability of NAD to induce T cell death at 24 h was assayed. Fig. 2 shows that 100 μM NAD increases annexin V staining in the PI-negative B6 T cell population from 5.8 to 81.4%, but there is no significant change in the percentage of PI staining cells. For BALB/c T cells, 100 μM NAD increases annexin V staining from 2.2 to 42.6% in the PI-negative population, but in contrast to B6 T cells, there is an increase from 48.9 to 82% PI staining cells. Therefore, NAD inhibits T cell proliferation and induces annexin V staining in both B6 and BALB/c T cells, but whereas NAD induces PI permeability in BALB/c T cells, this is not the case in B6 T cells.
FIGURE 2. Effects of increasing doses of NAD on B6 and BALB/c T cell viability. T cells were incubated for 24 h after adding NAD at the indicated concentrations. Cells were then stained with annexin V, PI, and anti-CD3 and analyzed by FACS. Values in the quadrants represent percentages. Upper panels show data from PI-negative cells stained for CD3 and annexin V. Lower panels show total cells analyzed for PI staining and forward light scatter (FSC). All data shown are from one of three repeat experiments.
ADPR inhibits cell proliferation and induces a death signal in B6 but not BALB/c T cells
Rat T cells hydrolyze NAD within 10 min of incubation (11) suggesting mouse T cells might do the same. Fig. 3 shows that hydrolysis to ADPR does occur but does not quite reach 50% by 2 h. Therefore, NAD hydrolysis by mouse T cells is slower than that by rat T cells. Also of note is that there are no differences between B6, BALB/c, and ART-2–/– T cells, supporting the notion that CD38 and not ART-2 is the principal NAD hydrolyzing cell surface enzyme (16). The finding that NAD is degraded during the time frame of our assays raises the possibility that NAD metabolites cause the effects ascribed to NAD.
FIGURE 3. Hydrolysis of NAD in the presence of T cells. B6, BALB/c, and ART-2–/– T cells were incubated in 200 μM [32P]NAD for the times indicated in complete RPMI 1640 without FBS. Supernatants were removed and analyzed by TLC. Plots show relative amounts of NAD and ADPR calculated from chromatographs on the left. The data shown are from one of three similar experiments.
To test whether ADPR or nicotinamide exert effects on T cells, both molecules were tested in T cell proliferation assays. Fig. 1C shows that proliferation of B6 and BALB/c T cells is inhibited by addition of 100 μM ADPR but not by nicotinamide. These results suggest that ADPR may be responsible for some of the effects seen with NAD, although 10-times higher concentrations of ADPR than of NAD are required to induce comparable effects.
Therefore, it was important to examine effects of ADPR on T cell survival. Fig. 4 shows that 1000 μM ADPR increases annexin V staining in the PI-negative B6 T cell population from 4.3 to 20.7% and PI staining from 44.7 to 78.6% at 24 h. This result is quite different from that seen with NAD in Fig. 2 in which there was a dramatic increase in annexin V staining but no increase in PI staining. In BALB/c T cells, ADPR fails to increase annexin V staining, and the PI staining cell population increases by only 12%. Again, this result contrasts to the dramatic effects of NAD in BALB/c T cells (Fig. 2). Therefore, it appears that while high concentrations of ADPR inhibit cell proliferation in both BALB/c and B6 T cells, only in B6 T cells does ADPR induce a significant increase in annexin V and PI staining. Thus, suppression of cell proliferation by ADPR does not correlate with cell death, and the effects of NAD on T cells cannot be solely explained by actions of its metabolite ADPR.
FIGURE 4. Effects of increasing doses of ADPR on B6 and BALB/c T cell viability. ADPR was added to T cells at the concentrations indicated. After 24 h cells were stained with annexin V, PI, and anti-CD3 and analyzed by FACS. Other conditions were as in Fig. 2. The data shown are from one of three repeat experiments.
NAD- but not ADPR-induced cell death is a rapid reaction
The finding that adding NAD and ADPR to B6 T cells is associated with different degrees of annexin V and PI staining at 24 h (Figs. 2 and 4) suggests that the two molecules promote cell death by different mechanisms. To examine this, the time course of cell death was examined in B6 T cells after the addition of either NAD or ADPR. Fig. 5 shows that 2 min after adding NAD, annexin V and PI staining reached a plateau. In contrast, attempts at demonstrating cell death by ADPR at these early time points were unsuccessful (data not shown). Therefore, NAD-induced cell death appears to be rapid while that induced by ADPR takes many hours. Results with BALB/c T cells confirm that NAD-induced cell death is exceedingly rapid as annexin V and PI staining plateau as early as 30 s after addition of NAD (Fig. 5). These results strongly suggest that NAD and ADPR induce cell death by different mechanisms and that B6 and BALB/c T cells differ in their sensitivity to these two molecules.
FIGURE 5. Time kinetics of NAD-induced cell death in B6 and BALB/c. NAD was added to T cells at a concentration of 500 μM. After the indicated times, the cells were stained with annexin V, PI, and anti-CD3 and analyzed by FACS. Other conditions were as in Fig. 2. The data shown are from one of three repeat experiments.
Features of apoptosis characterize slow but not rapid cell death induced by NAD
In light of the above results, it was of interest to identify mechanisms involved in inducing rapid and slow cell death. Cell death often proceeds by apoptotic mechanisms, and a hallmark of apoptosis is the activation of caspases. Therefore, we tested whether inhibition of caspases interferes with rapid or slow induction of cell death. B6 T cells were first incubated with caspase inhibitor Z-VADfmk and then different concentrations of NAD were added. Fig. 6A shows that after 24 h, compared with no addition, adding NAD increases the PI-negative, annexin V-positive cell population from 5.1 to 81.2%, and that Z-VADfmk reduces staining to 26.7%. Therefore, the mechanism by which NAD induces slow cell death appears to involve the action of caspases. Because slow cell death is also seen with ADPR, we assayed effects of Z-VADfmk on ADPR-induced cell death. Fig. 6A shows ADPR increases the PI-negative annexin V-positive population from 5.1 to 19.9%, which is suppressed by the inhibitor. ADPR also increases the PI staining population, i.e., from 45.6 to 78%, which is decreased to 52.2% by Z-VADfmk (Fig. 6A). These results suggest that the slow induction of cell death by both NAD and ADPR involves caspases.
FIGURE 6. Effect of Z-VADfmk on NAD- and ADPR-induced cell death. A, NAD and ADPR were added to B6 T cells at 500 μM and cells analyzed by FACS after 24 h. Where indicated cells were first incubated with 50 μM Z-VADfmk for 1 h before addition of NAD or ADPR. Cells were stained with annexin V, PI, and anti-CD3 and analyzed by FACS. Other conditions were as in Fig. 2. B, Effect of Z-VADfmk on NAD-induced cell death in B6 and BALB/c T cells in the 2- and 24-h assays. NAD was added to T cells at 500 μM, and the cells were analyzed after 2 or 24 h. Data are from PI-negative cells stained for CD3 and annexin V. Other conditions were as in A. All experiments were repeated three times, and one representative experiment is shown.
Therefore, we examined whether caspases are also involved in the rapid induction of cell death. Fig. 6B shows that annexin V staining in the PI-negative B6 T cell population increases from 4.9 to 13.7% within 2 h of NAD addition and that Z-VADfmk does not inhibit this effect. Therefore, annexin V staining induced during the first 2 h of incubation appears to be caspase independent. BALB/c T cells show entirely concordant results (Fig. 6B).
To further test the hypothesis that apoptosis is not involved in the rapid induction of death, arising from the addition of NAD, T cells were assayed for DNA fragmentation. Fig. 7 shows that 2 h after the addition of NAD or ADPR there is no increase in DNA fragmentation assayed by TUNEL staining in either B6 or BALB/c T cells. During this time period, annexin V staining increased from 3.5 to 19.7% in B6 cells and from 3.2 to 58.9% in BALB/c T cells. Therefore, cell death in the first 2 h after addition of NAD does not involve DNA fragmentation. In contrast, TUNEL staining increased significantly in BALB/c and B6 T cells 24 h after the addition of NAD (Fig. 7). Adding ADPR instead of NAD produced similar but less dramatic effects after 24 h (Fig. 7). Therefore, both NAD and ADPR induce slow cell death by an apoptotic pathway, associated with DNA fragmentation.
FIGURE 7. Induction of DNA fragmentation by NAD and ADPR in B6 and BALB/c cells. NAD or ADPR was added to T cells at 500 μM. After 24 h, the cells were stained with anti-CD3, annexin V, PI, and TUNEL (BrdU). Shown are results from CD3-gated, BrdU-stained cells. Annexin V staining is shown for the PI-negative, CD3-positive cell population. The data shown are from one of three repeat experiments.
ART-2 is required for rapid but not slow cell death induced by NAD
The finding that NAD but not ADPR induces rapid cell death is consistent with the notion that ADP-ribosylation is involved. However, because NAD also induces slow cell death, the question arises whether ADP-ribosylation contributes to the slow induction of cell death.
To find out, cell death was examined in T cells from wild-type NOD and NOD ART-2 gene-deleted mice. Fig. 8A shows that 2 h after adding NAD to NOD ART-2–/– T cells, there is no increase in annexin V staining shown in the dot plots or PI staining represented by percentages below the plots. In contrast, NAD effectively induces staining in T cells from wild-type NOD mice. Therefore, rapid cell death is not observed in T cells lacking ART-2, consistent with the requirement of ADP-ribosylation.
FIGURE 8. Effects of NAD and ADPR on T cells from wild-type NOD and ART-2–/– NOD mice. A, Effect of Z-VAD fmk on NAD- and ADPR-induced cell death in wild-type and ART-2–/– T cells in the 2- and 24-h assays. NAD or ADPR was added to T cells at 500 μM and analyzed after 2 or 24 h. Where indicated, cells were first incubated with 50 μM Z-VADfmk for 1 h before addition of NAD or ADPR. Cells were stained with annexin V, PI, and anti-CD3 and analyzed by FACS. Dot plots show data for PI-negative cells stained for CD3 and annexin V. Note that the percentages of PI-positive T cells from each analysis and condition are shown as percentages under each panel. B, Induction of DNA fragmentation by NAD and ADPR in ART2–/– T cells. NAD or ADPR were added to T cells at 500 μM. After 24 h, the cells were stained with anti-CD3, annexin V, PI, and TUNEL (BrdU). Shown are results from CD3-gated, BrdU-stained cells. Annexin V staining is shown for the PI-negative, CD3-positive cell population. The data shown are from one of three repeat experiments.
In contrast, 24 h after the addition of NAD to NOD ART-2–/– T cells there is an increase in annexin V staining from 5.7 to 45%, which is suppressed to 11.4% by Z-VADfmk, seen in the dot plots. Although ADPR causes only a small increase of annexin V staining (5.7 to 10.2%), there is an increase in the PI staining cell population from 37.6 to 61.8% as indicated by the percentages under the individual panels. Both effects are decreased by Z-VADfmk. Similar results are seen in T cells from NOD control mice, but note that NAD-induced effects are more pronounced in the wild type compared with the ART-2–/– T cells. Thus, there is a slow induction of cell death that does not require a functional ART-2 enzyme. To further test this conclusion, DNA fragmentation was assayed in ART-2–/– T cells. Fig. 8B shows that both NAD and ADPR induce a moderate increase in TUNEL staining at 24 h. However, the two molecules likely act by different mechanisms, because ADPR but not NAD induces a substantial increase in PI staining at the 24-h time point (Fig. 8A).
Purinergic receptor P2X7 is required for NAD to induce rapid but not slow cell death
The finding that NAD degradation product ADPR mimics some of the effects induced by NAD, points to the involvement of purinergic receptors. Among the family of purinergic receptors, P2X7 is widely expressed on cells of the hemopoietic cell lineage (17, 18, 19, 20). Recent data also suggests that P2X7, whose well-established ligand is ATP, expresses low affinity for ADP (21). Moreover, inhibitors of P2X7 receptors can block induction of cell death by NAD (10). Taken together, these observations make P2X7 receptors a prime candidate to mediate effects of NAD and ADPR. To examine whether the rapid or slow mechanism of cell death, induced by NAD or ADPR, involve the P2X7 receptor, use was made of the finding that KN-62 inhibits within minutes the signaling function of this receptor (15, 22). BALB/c T cells were incubated with KN-62 for 10 min, followed by addition of NAD; 15 min later, T cells were stained for annexin V and PI. Fig. 9A shows that KN-62 completely inhibits NAD-induced annexin V and PI staining in this short-term assay. Similar results were obtained with B6 T cells (data not shown). However, we were not able to test KN62 in the slow induction of death because of its toxicity (data not shown).
FIGURE 9. A, Effect of KN-62 on NAD-induced cell death in BALB/c T cells. NAD was added to T cells at 500 μM. After 15 min, the cells were stained and subjected to FACS analysis. Where indicated, cells were first incubated with 20 μM KN-62 for 10 min before addition of NAD. Cells were stained with annexin V, PI, and anti-CD3 and analyzed by FACS. Values in the quadrants represent percentages. Panels show data from PI-negative cells stained for CD3 and annexin V, as well as total cells analyzed for PI staining and forward light scatter (FSC). B and C, Induction of cell death by NAD and ADPR in T cells lacking P2X7 receptors. NAD or ADPR were added to T cells from B6 and P2X7–/– mice at 500 μM. After 2 and 24 h, cells were stained with CD3, annexin V, and PI and analyzed by FACS. Annexin V staining is shown for the PI-negative, CD3-positive cell population. The data shown are from one of three repeat experiments.
Therefore, we examined T cells from P2X7 receptor gene-deleted B6 mice (12). Fig. 9, B and C, shows that 2 h after addition of NAD, B6 (Fig. 9B) but not P2X7–/– (Fig. 9C) T cells displayed increased annexin V and PI staining. At 24 h, significant annexin V and PI staining is seen in B6 cells (69.6%) and somewhat lower staining is seen in P2X7–/– T cells (51%). This result shows that deleting expression of P2X7 receptors prevents NAD from rapidly inducing cell death. In contrast, NAD still induced the slow cell death in the P2X7–/– T cells. Results with ADPR are less striking. Twenty-four hours after adding ADPR, annexin V staining had increased from 8.2 to 15.2% for B6 cells, but was only 7.4% in P2X7–/– T cells. Thus, a functional P2X7 receptor is required for NAD to rapidly induce cell death but not slow death. In contrast, the P2X7 receptor may play a role in the mechanism by which ADPR induces slow cell death.
P2X7 receptor function is not impaired in etheno-NAD-treated cells
Two hypotheses arise to explain how cell surface ADP-ribosylation induced by NAD leads to rapid cell death. One is that the ADP-ribosylation of a protein modulates its function. The other is that ADP-ribosyl groups serve as ligands for a receptor. To test these hypotheses, we assessed effects of ADP-ribosylation on P2X7 receptor-triggered cell death by its ligand ATP. As previously reported (13), incubating T cells with etheno-NAD leads to etheno-ADP-ribosylation (Fig. 1, A and B) but not cell death. In fact, etheno-ADP-ribosylation renders T cells resistant to actions of NAD (10), presumably by filling the sites that are ADP-ribosylated upon the addition of NAD. Fig. 10 shows that while incubation of BALB/c T cells with etheno-NAD does not induce annexin V or PI staining, it does block the action of a subsequent addition of NAD. In contrast, etheno-NAD only slightly blocks rapid induction of annexin V by ATP and PI staining. These results are consistent with the notion that cell surface ADP-ribosyl groups act as ligands for P2X7 receptors, but not with the possibility that ADP-ribosylation acts by modifying protein function.
FIGURE 10. Effects of ATP on BALB/c T cells after incubation with etheno-NAD. Etheno-NAD was added to T cells at 300 μM. After 30 min, cells were washed and then incubated with ATP or NAD at the indicated concentrations. After 90 min, the cells were stained with annexin V, PI, and anti-CD3 and analyzed by FACS. Values in the quadrants represent percentages. Data from PI-negative cells stained for CD3 and annexin V are shown. The data is from one of three repeat experiments.
Discussion
The observation that murine T cells express cell surface ADP-ribosyltransferase activity and that by providing the enzyme with the substrate NAD leads to modification of cell surface proteins, which is associated with inhibition of T cell functions and induction of cell death (1, 2, 3, 4), has raised intriguing questions as to the physiological relevance of this reaction. An important issue is whether NAD concentrations that trigger cell death can be reached in vivo. Although intracellular NAD approaches 1000 μM, levels in blood are only 0.13 μM (23, 24). As shown here and elsewhere (1, 4), concentrations as low as 1 μM NAD can exert effects on T cells. Therefore, NAD levels one-thousandth of those inside cells and 10 times higher then those in circulation can mediate suppressive effects. Moreover, this report demonstrates that NAD need be present for only seconds for effects to be evident. Thus, effective NAD levels could well be achieved during tissue injury or by constitutive release from live cells.
Another question is whether induction of death arises from NAD itself or its breakdown products. The results above confirm the observation (11) that T cells substantially hydrolyze NAD to ADPR, and that upon prolonged incubation ADPR exerts effects similar but not identical to those induced by NAD. It should also be noted that ADPR is metabolized to adenosine (25), which could act on T cells via action on purinergic receptors. Indeed effects of adenosine on T cells via action on A2A receptors are well documented in the literature (for review see Ref.26). Therefore, effects induced by the addition of NAD, particularly in incubations lasting 24 h, may very well result from the combined effects of NAD and its catabolic derivatives.
This report demonstrates that NAD can induce annexin V and PI staining in T cells by different mechanisms that lead to either rapid or slow cell death. Rapid cell death is not inhibited by caspase inhibitor Z-VADfmk, and therefore, is reminiscent of necrosis. Slow cell death is associated with DNA fragmentation and is inhibited by caspase inhibitor Z-VADfmk, and therefore, has features of apoptosis. It should be noted that while both NAD and ADPR slowly induce apoptotic death, they do not appear to act by identical mechanisms. This again may be due to action of NAD metabolites on purinergic receptors such as A2A receptors or others yet to be identified (26)
It is interesting that adding NAD to rat T cells leads to autoADP-ribosylation of cell surface ART-2 but apparently no other proteins (27, 28). Although adding either NAD or ADPR to rat T cells inhibits their proliferation, NAD does not induce a rapid increase in annexin V staining (28). Therefore, it appears that autoADP-ribosylation of cell surface ART-2 is not sufficient to induce rapid T cell death.
An enigmatic question is how NAD inhibits T cell functions and induces cell death. Initially, it appeared that effects observed after adding NAD to T cells correlated with cell surface ADP-ribosylation (2, 4). Subsequent data showed this correlation is not a tight one. B6 T cells undergo substantially more ADP-ribosylation than BALB/c T cells, yet NAD is less effective in promoting rapid cell death in B6 cells. An important difference between BALB/c and B6 mice is the differential expression of the two ART-2 genes ART-2a and ART-2b, which could be responsible for this effect. BALB/c mice express both ART-2a and ART-2b, whereas B6 mice express ART-2b only, because of a mutation in ART-2a (4). This raises the possibility that ADP-ribosylation by the ART-2a coded enzyme is more efficient in cell death induction then that by the ART-2b enzyme. However, this does not appear to be the case because NZW mice expressing only ART-2a possess T cells even less sensitive to NAD than B6 mice T cells (4). Therefore, it was important to demonstrate that ADP-ribosylation is critical for the induction of cell death. We show that deletion of the ART-2 genes in NOD mice blocks the rapid induction of cell death by NAD, consistent with results generated previously in a different mouse strain (7).
We present experimental evidence that deleting expression of the P2X7 gene blocks rapid cell death induction by NAD. These data prompt a plausible explanation for the apparent lack of correlation between the extent of T cell ADP-ribosylation and induction of rapid cell death in T cells from B6 and BALB/c mice. BALB/c T cells express the wild-type P2X7 receptor, which mediates efficient and rapid cell death upon binding ATP. In contrast, the B6 P2X7 receptor has a mutation in the TNFR1 deathlike domain resulting in much less efficient induction of cell death (20). Therefore, it appears that reduced induction of cell death by NAD in B6 compared with BALB/c T cells arises from the mutation altering the function of the P2X7 receptor in B6 mice.
Fig. 11 shows a model for the role of P2X7 receptors in mediating the effects arising from addition of NAD to cultured T cells. Using NAD, ART-2 attaches ADP-ribosyl groups to a cell surface protein near the P2X7 receptor or to the receptor itself. The ligand-binding site of the receptor engages the ADP-ribosyl groups, rapidly inducing cell death by pore formation. The data presented show that the death signal is induced within 30 s of NAD contact. Given the evidence that induction of the death signal involves activation of P2X7 receptors, it is likely that the formation of membrane pores is preceded by calcium fluxes. Indeed, incubation of mouse T cells with NAD induces calcium fluxes (10). Similar results were recently reported with human monocytes in which both NAD and ADPR induce calcium fluxes, but it was not determined which receptors may be involved (29). But, as previously demonstrated, the P2X7 receptor is also capable of inducing a slower, caspase-dependent apoptotic cell death (30, 31). Attachment of etheno-ADP-ribosyl groups fails to trigger rapid death via P2X7 activation due to the structural specificity of the P2X7 binding site, and by filling potential sites for ADP-ribosylations, etheno-NAD blocks the action of subsequently added NAD. However, after etheno-ADP-ribosylation, P2X7 receptors remain functional as ATP can still induce rapid cell death. In addition, metabolites of NAD and ADPR may react with other receptors that lead to slow cell death. Exposure of phosphatidyl serine, and thus binding of annexin V, characterizes the induction of both rapid and slow cell death.
FIGURE 11. A model for ecto-NAD-induced effects on T cells. Using NAD as a substrate, ART-2 attaches ADP-ribosyl groups to cell surface protein(s) near P2X7 receptors. Binding of covalently attached ADP-ribosyl groups to P2X7 induces cell death by rapid pore formation and possibly slower caspase activation. Attachment of etheno-ADP-ribosyl groups blocks sites that otherwise would be ADP-ribosylated upon addition of NAD. Etheno-ADP-ribosyl groups do not trigger P2X7 signaling and do not interfere with ATP signaling. NAD also undergoes hydrolysis to ADPR, which itself or when further catabolized may provide ligands for other purinergic receptors, inducing slower caspase-dependent apoptosis. Both the rapid and slow pathways involve exposure of phosphatidyl serine.
It is important to note that for ATP to trigger P2X7 signaling, concentrations are required that are considerably higher than those needed for NAD (32). Moreover, adding NAD to cells leads to ADP-ribosylations within seconds, which persist after NAD is removed (5, 10). Given that intracellular concentrations of NAD and ATP are similar (23, 33), it becomes plausible that NAD released during trauma may be at least as important as ATP in regulating T cell functions via P2X7 receptors.
The question arises as to whether the model proposed above explains observations in earlier reports. Consistent with our model may be the observation that depleting rats or mice of T cells expressing ART-2 accelerates development of autoimmune diabetes and lupus erythematosus (34, 35). This could be explained by a mechanism in which NAD released from cells controls effector T cells. However, we have observed that deletion of the tandem ART-2 genes in the NOD mouse strain used in the above experiments (36) has no effect on diabetes progression (data not shown), even though lymphocytes from the parent NOD strain are particularly sensitive to signaling through P2X7 receptors (37). Taken together, these observations indicate the need to further study the role of ecto-NAD and its metabolites in the induction of cell death.
Acknowledgments
We thank Dr. R. Santella for the gift of Ab 1G4 and Dr. C. Gabel for generously providing the P2X7 gene-deleted mice.
Footnotes
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.
1 This work was supported by Public Health Service Grants AI 40038 and AI 43954.
2 Address correspondence and reprint requests to Dr. Gunther Dennert at the current address: University of Southern California/Norris Comprehensive Cancer Center, P.O. Box 33800, 1441 Eastlake Avenue, M/S 73, Los Angeles, CA 90033-0800. E-mail address: dennert@usc.edu
3 Abbreviations used in this paper: ART, ADP-ribosyltransferase; NWNA, nylon wool nonadherent; ADPR, ADP-ribose; PI, propidium iodide.
Received for publication September 21, 2004. Accepted for publication November 18, 2004.
References
Wang, J., E. Nemoto, A. Y. Kots, H. R. Kaslow, G. Dennert. 1994. Regulation of cytotoxic T cells by ecto-nicotinamide adenine dinucleotide (NAD) correlates with cell surface GPI-anchored/arginine ADP-ribosyltransferase. J. Immunol. 153:4048.
Liu, Z. X., O. Azhipa, S. Okamoto, S. Govindarajan, G. Dennert. 2001. Extracellular nicotinamide adenine dinucleotide induces T cell apoptosis in vivo and in vitro. J. Immunol. 167:4942.
Han, M. K., Y. S. Cho, Y. S. Kim, C. Y. Yim, U. H. Kim. 2000. Interaction of two classes of ADP-ribose transfer reactions in immune signaling. J. Biol. Chem. 275:20799.
Adriouch, S., W. Ohlrogge, F. Haag, F. Koch-Nolte, M. Seman. 2001. Rapid induction of naive T cell apoptosis by ecto-nicotinamide adenine dinucleotide: requirement for mono(ADP-ribosyl)transferase 2 and a downstream effector. J. Immunol. 167:196.
Nemoto, E., Y. Yu, G. Dennert. 1996. Cell surface ADP-ribosyltransferase regulates lymphocyte function-associated molecule-1 (LFA-1) function in T cells. J. Immunol. 157:3341.
Kahl, S., M. Nissen, R. Girisch, T. Duffy, E. H. Leiter, F. Haag, F. Koch-Nolte. 2000. Metalloprotease-mediated shedding of enzymatically active mouse ecto-ADP-ribosyltransferase ART2.2 upon T cell activation. J. Immunol. 165:4463.
Ohlrogge, W., F. Haag, J. Lohler, M. Seman, D. R. Littman, N. Killeen, F. Koch-Nolte. 2002. Generation and characterization of ecto-ADP-ribosyltransferase ART2.1/ART2.2-deficient mice. Mol. Cell. Biol. 22:7535.
Liu, Z. X., Y. Yu, G. Dennert. 1999. A cell surface ADP-ribosyltransferase modulates T cell receptor association and signaling. J. Biol. Chem. 274:17399.
Okamoto, S., O. Azhipa, Y. Yu, E. Russo, G. Dennert. 1998. Expression of ADP-ribosyltransferase on normal T lymphocytes and effects of nicotinamide adenine dinucleotide on their function. J. Immunol. 160:4190
Seman, M., S. Adriouch, F. Scheuplein, C. Krebs, D. Freese, G. Glowacki, P. Deterre, F. Haag, F. Koch-Nolte. 2003. NAD-induced T cell death: ADP-ribosylation of cell surface proteins by ART2 activates the cytolytic P2X7 purinoceptor. Immunity 19:571.
Bortell, R., J. Moss, R. C. McKenna, M. R. Rigby, D. Niedzwiecki, L. A. Stevens, W. A. Patton, J. P. Mordes, D. L. Greiner, A. A. Rossini. 2001. Nicotinamide adenine dinucleotide (NAD) and its metabolites inhibit T lymphocyte proliferation: role of cell surface NAD glycohydrolase and pyrophosphatase activities. J. Immunol. 167:2049.
Solle, M., J. Labasi, D. G. Perregaux, E. Stam, N. Petrushova, B. H. Koller, R. J. Griffiths, C. A. Gabel. 2001. Altered cytokine production in mice lacking P2X(7) receptors. J. Biol. Chem. 276:125.
Krebs, C., W. Koestner, M. Nissen, V. Welge, I. Parusel, F. Malavasi, E. H. Leiter, R. M. Santella, F. Haag, F. Koch-Nolte. 2003. Flow cytometric and immunoblot assays for cell surface ADP-ribosylation using a monoclonal antibody specific for ethenoadenosine. Anal. Biochem. 314:108.
Matsuoka, S., H. Tsurui, M. Abe, K. Terashima, K. Nakamura, Y. Hamano, M. Ohtsuji, N. Honma, I. Serizawa, Y. Ishii, et al 2003. A monoclonal antibody to the 2 domain of murine major histocompatibility complex class I that specifically kills activated lymphocytes and blocks liver damage in the concanavalin A hepatitis model. J. Exp. Med. 198:497.
Humphreys, B. D., C. Virginio, A. Surprenant, J. Rice, G. R. Dubyak. 1998. Isoquinolines as antagonists of the P2X7 nucleotide receptor: high selectivity for the human versus rat receptor homologues. Mol. Pharmacol. 54:22.
Howard, M., J. C. Grimaldi, J. F. Bazan, F. E. Lund, L. Santos-Argumedo, R. M. Parkhouse, T. F. Walseth, H. C. Lee. 1993. Formation and hydrolysis of cyclic ADP-ribose catalyzed by lymphocyte antigen CD38. Science 262:1056.
Mutini, C., S. Falzoni, D. Ferrari, P. Chiozzi, A. Morelli, O. R. Baricordi, G. Collo, P. Ricciardi-Castagnoli, F. Di Virgilio. 1999. Mouse dendritic cells express the P2X7 purinergic receptor: characterization and possible participation in antigen presentation. J. Immunol. 163:1958.
Collo, G., S. Neidhart, E. Kawashima, M. Kosco-Vilbois, R. A. North, G. Buell. 1997. Tissue distribution of the P2X7 receptor. Neuropharmacology 36:1277
Mehta, V. B., J. Hart, M. D. Wewers. 2001. ATP-stimulated release of interleukin (IL)-1 and IL-18 requires priming by lipopolysaccharide and is independent of caspase-1 cleavage. J. Biol. Chem. 276:3820.
Adriouch, S., C. Dox, V. Welge, M. Seman, F. Koch-Nolte, F. Haag. 2002. Cutting edge: a natural P451L mutation in the cytoplasmic domain impairs the function of the mouse P2X7 receptor. J. Immunol. 169:4108.
Chakfe, Y., R. Seguin, J. P. Antel, C. Morissette, D. Malo, D. Henderson, P. Seguela. 2002. ADP and AMP induce interleukin-1 release from microglial cells through activation of ATP-primed P2X7 receptor channels. J. Neurosci. 22:3061.
Gargett, C. E., J. S. Wiley. 1997. The isoquinoline derivative KN-62 a potent antagonist of the P2Z-receptor of human lymphocytes. Br. J. Pharmacol. 120:1483.
Jacobson, E. L., M. K. Jacobson. 1997. Tissue NAD as a biochemical measure of niacin status in humans. Methods Enzymol. 280:221.
Kim, U. H., M. K. Kim, J. S. Kim, M. K. Han, B. H. Park, H. R. Kim. 1993. Purification and characterization of NAD glycohydrolase from rabbit erythrocytes. Arch. Biochem. Biophys. 305:147.(Hiroki Kawamura, Fred Asw)
Adding NAD to murine T lymphocytes inhibits their functions and induces annexin V binding. This report shows that NAD induces cell death in a subset of T cells within seconds whereas others do not die until many hours later. Low NAD concentrations (<10 μM) suffice to trigger rapid cell death, which is associated with annexin V binding and membrane pore formation, is not blocked by the caspase inhibitor Z-VADfmk, and requires functional P2X7 receptors. The slower induction of death requires higher NAD concentrations (>100 μM), is blocked by caspase inhibitor Z-VADfmk, is associated with DNA fragmentation, and does not require P2X7 receptors. T cells degrade NAD to ADP-ribose (ADPR), and adding ADPR to T cells leads to slow but not rapid cell death. NAD but not ADPR provides the substrate for ADP-ribosyltransferase (ART-2)-mediated attachment of ADP-ribosyl groups to cell surface proteins; expression of ART-2 is required for NAD to trigger rapid but not slow cell death. These results support the hypothesis that cell surface ART-2 uses NAD but not ADPR to attach ADP-ribosyl groups to the cell surface, and that these groups act as ligands for P2X7 receptors that then induce rapid cell death. Adding either NAD or ADPR also triggers a different set of mechanisms, not requiring ART-2 or P2X7 receptors that more slowly induce cell death.
Introduction
The addition of NAD to murine T lymphocytes inhibits their functions (1) and induces annexin V binding (2, 3, 4). Treating naive T cells with NAD before injection into semiallogeneic mice blocks induction of graft-vs-host disease (2). Treating T cells before injection into syngeneic mice blocks their migration to the spleen and directs them to the liver where they appear to die (2). Injecting NAD directly into mice increases the number of apoptotic T cells in the liver (2). Taken together, these results indicate extracellular NAD (ecto-NAD) could play a role in immune regulation.
Several hypotheses have emerged to explain actions of ecto-NAD on T cells. One hypothesis is that it serves as a substrate for an arginine-specific cell surface ADP-ribosyltransferase (ART-2),3 which by ADP-ribosylating cell surface proteins causes the observed effects (1). Several observations support this hypothesis; anti-CD3 activation of murine T cells releases ART-2 and causes the cells to become resistant to NAD (5, 6), and using recombinant means, deletion of murine T cell ART-2 isoforms (ART2a and ART2b) inhibits cell surface ADP-ribosylation (7). Several of the proteins incorporating ADP-ribosyl groups have been identified, including LFA-1, CD27, CD43, CD45, and CD8 (8, 9), raising questions as to whether modification of one or more of these proteins induces cell death. It has been proposed that effects of ecto-NAD may arise from an activating ADP-ribosylation of purinergic receptors (10). It has also been proposed that ecto-NAD leads to an inactivating ADP-ribosylation of CD38, a cell surface NAD-glycohydrolase that can synthesize cyclic ADP-ribose (ADPR) (3). Yet another hypothesis arises from the observation that ecto-NAD inhibits mitogen-stimulated proliferation of rat T cells, which express little or no ART-2, and that, for rat T cells, adding ADPR is as effective as adding NAD. Because rat T cells degrade NAD to ADPR, the hypothesis arose that ADP-ribosylation may not be involved in the effects arising from the addition of NAD to cultured T cells but rather that NAD breakdown products act as ligands for purinergic receptors (11).
The data reported herein prompt a unifying model consistent with the above observations. The data demonstrate that adding NAD to T cells induces both rapid and slow death involving different sets of mechanisms. Rapid death involves pore formation and signaling through P2X7 receptors while slower death requires caspase activation but not P2X7 receptors. It is suggested that P2X7 engagement of ADP-ribosyl groups, attached by action of ART-2 to cell surface proteins, triggers receptor function and rapid cell death. This mechanism may provide a means by which NAD released during trauma controls T cell functions.
Materials and Methods
Mouse strains, T cell purification, and flow cytometric analysis
Pathogen-free female C57BL/6 (B6), and BALB/c mice, 6–8 wk of age were obtained from The Jackson Laboratory. NOD.129S4(B6)-Art2aTm1FknArt2bTm1Fkn (henceforth denoted as ART-2–/– mice) were produced at The Jackson Laboratory by outcross to a previously described B6 stock carrying targeted mutations in the tandem genes encoding the ART-2.1 and ART-2.2 ectoenzymes (7). Following nine backcrosses to NOD/Lt (N10) and verifying presence of NOD alleles at all known "Idd" loci, intercrosses were initiated, and ART2–/– homozygous breeding stock were received and bred at the University of Southern California animal facility (Los Angeles, CA). B6 P2X7–/– mice were kindly provided by Dr. C. Gabel (Ann Arbor, MI) and Pfizer and were bred at the University of Southern California animal facility (12). T cells were purified from spleen cells by nylon wool nonadherent (NWNA) cells and cultured in complete RPMI 1640 medium, containing 10% FBS (2). By FACS analysis, 85–90% of the NWNA cells were CD3 positive. For FACS analysis, T cells were preincubated with anti-mouse CD16/CD32 (2.4G2) mAb from BD Biosciences to block FcRs, and then incubated with various mAbs for 30 min at 4°C (2). Phenotypic analysis was performed using PE Cy5-conjugated anti-mouse CD3 (145-2C11) (BD Biosciences). To monitor induction of cell death, cells were stained with the Annexin V-FITC Apoptosis Detection Kit I (BD Biosciences). To demonstrate DNA fragmentation, cells were assayed using the TUNEL APO-Direct Apoptosis Assay kit (BD Biosciences). To quantitate cell surface ADP-ribosylation, NWNA cells (4 x 106 cells/ml in RPMI 1640) were incubated with etheno-NAD (Sigma-Aldrich), followed by incubation with etheno-ADPR-specific Ab 1G4, kindly provided by Dr. R. Santella (Mailma School of Public Health of Columbia University, New York, NY) (13) and a FITC goat anti-mouse Ig (BD Biosciences). FACS analysis was performed on a FACSCalibur (BD Biosciences).
Quantification of NAD metabolites
NWNA B6, BALB/c, and ART2–/– T cells (2 x 106) were incubated in 50 μl of serum-free PBS containing 200 μM NAD (Sigma-Aldrich) and 0.1 μCi [32P]NAD (Amersham) at 37°C. After centrifugation, supernatant from each sample (2 μl) was removed and applied to a 20 x 20 cm PEI-cellulose F glass-backed TLC plate (EMD Chemicals). TLC was developed in 0.3 M LiCl (Sigma-Aldrich) and exposed to a phosphor screen (Amersham) for 3 h at 25°C. NAD and ADPR were visualized using a PhosphorImager 445SI (Amersham), and radioactivity was quantified by ImageQuant 5.0 (Amersham) (11).
Western blot analyses
NWNA cells were incubated in RPMI 1640 with or without etheno-NAD (300 μM) for 30 min at 37°C. After washing twice in ice cold PBS, cells were re-suspended in lysis buffer (108 cells/ml PBS containing 1% Igepal, 1 mM AEBSF) and incubated for 30 min at 4°C. Insoluble material was removed by centrifugation (15 min, 10,000 x g), and supernatants were mixed with sample buffer, heated for 8 min at 95°C, and then loaded on SDS-PAGE gels. After electrophoresis, proteins were transferred onto PVDF membrane (Bio-Rad). Membranes were blocked with blocking buffer (LI-COR) for 1 h at room temperature. Blots were then incubated with 1G4 Ab (13) (ascites 1/500 dilution in TBS, 1% BSA, 0.05% Tween 20) for 2 h at room temperature. Membranes were washed with 0.05% Tween 20 in TBS and incubated with a 1/5000 dilution of Alexis Fluor 680-conjugated secondary anti-mouse IgG Abs (Molecular Probes) for 1 h. Protein bands were visualized by an infrared image system (LI-COR).
Cell proliferation and cell death assays and inhibition of caspases and P2X7 receptor function
To assay cell proliferation, NWNA cells were cultured on anti-CD3-coated tissue culture plates (2). Plates were incubated with a 1/1000 dilution of anti-CD3 Ab (500AA2 ascites) overnight. After washing the plates, T cells (5 x 105/well) were added in complete RPMI 1640 medium containing 10% FBS, and incubation was continued for 48 h. [3H]Thymidine (0.5 μCi/well) was added during the last 18 h of culture.
To assay induction of cell death, NWNA T cells were incubated with or without NAD, etheno-NAD, ADPR, or ATP (Sigma-Aldrich) for various times and assayed for propidium iodide (PI) uptake, annexin V, or TUNEL staining. To assay short-term effects of NAD (5 min), a 10-fold excess ice-cold PBS was added to the cell suspension to dilute NAD. Following centrifugation, cells were washed twice in ice-cold PBS before culture in complete RPMI 1640 medium. Assays up to 2 h were done in complete RPMI 1640 medium without FBS, while those for 24 h contained 10% FBS. Caspase inhibitor Z-VADfmk (BD Biosciences) was used at 50 μM and added to cell cultures 1 h before addition of NAD (14). To inhibit P2X7 receptors, 20 μM KN-62 (Sigma-Aldrich) dissolved in 0.01% DMSO were added 10 min before addition of NAD or ATP (15).
Results
NAD inhibits cell proliferation and induces a death signal in both BALB/c and B6 T cells
Adding NAD to either B6 or BALB/c T cells leads to cell surface ADP-ribosylation (2, 4) and increases annexin V staining, but there are significant differences. B6 T cells undergo more extensive cell surface ADP-ribosylation than BALB/c T cells (1, 4), but a higher proportion of BALB/c T cells is induced for cell death (4). These differences become evident when purified T cells are incubated with ART-2 substrate etheno-NAD and then stained with etheno-ADPR-specific Ab 1G4 (13) (Fig. 1A). BALB/c T cells show staining when compared with cells from ART-2 gene-deleted mice (7), but it is much less pronounced then that of B6 T cells. Concordant results are seen in immunoblots. Cell extracts from BALB/c show barely detectable bands, whereas those from B6 T cells show strong bands (Fig. 1B). These differences between B6 and BALB/c T cells provide an experimental system to study the pathways responsible for NAD-induced cell death.
FIGURE 1. ADP-ribosylation of T cells and effects of NAD, ADPR, and nicotinamide on T cell proliferation. A, ADP-ribosylation of B6, BALB/c, and ART-2–/– T cells. T cells were incubated with 300 μM etheno-NAD for 30 min, stained with Ab 1G4, and analyzed by FACS. B, Cell lysates were prepared from etheno-NAD treated cells separated by PAGE and analyzed by immunoblotting using 1G4 Ab. C, Inhibition of T cell proliferation by NAD, ADPR, and nicotinamide. B6 and BALB/c T cells were cultured on anti-CD3-coated tissue culture plates with the indicated concentrations of NAD, ADPR, and nicotinamide for 48 h. [3H]TdR was added for the last 18 h of incubation, and incorporation of 3H was determined. Error bars indicate SD. All data shown are from one of three repeat experiments.
Effects of NAD on cell proliferation were compared with T cells from B6 and BALB/c mice. Fig. 1C shows that anti-CD3 stimulated [3H]TdR incorporation into B6 and BALB/c T cells is almost completely inhibited by adding NAD at concentrations between 10 and 1000 μM. Therefore, effects of NAD on T cell proliferation are not proportional to the extent of cell surface ADP-ribosylation.
Next, the ability of NAD to induce T cell death at 24 h was assayed. Fig. 2 shows that 100 μM NAD increases annexin V staining in the PI-negative B6 T cell population from 5.8 to 81.4%, but there is no significant change in the percentage of PI staining cells. For BALB/c T cells, 100 μM NAD increases annexin V staining from 2.2 to 42.6% in the PI-negative population, but in contrast to B6 T cells, there is an increase from 48.9 to 82% PI staining cells. Therefore, NAD inhibits T cell proliferation and induces annexin V staining in both B6 and BALB/c T cells, but whereas NAD induces PI permeability in BALB/c T cells, this is not the case in B6 T cells.
FIGURE 2. Effects of increasing doses of NAD on B6 and BALB/c T cell viability. T cells were incubated for 24 h after adding NAD at the indicated concentrations. Cells were then stained with annexin V, PI, and anti-CD3 and analyzed by FACS. Values in the quadrants represent percentages. Upper panels show data from PI-negative cells stained for CD3 and annexin V. Lower panels show total cells analyzed for PI staining and forward light scatter (FSC). All data shown are from one of three repeat experiments.
ADPR inhibits cell proliferation and induces a death signal in B6 but not BALB/c T cells
Rat T cells hydrolyze NAD within 10 min of incubation (11) suggesting mouse T cells might do the same. Fig. 3 shows that hydrolysis to ADPR does occur but does not quite reach 50% by 2 h. Therefore, NAD hydrolysis by mouse T cells is slower than that by rat T cells. Also of note is that there are no differences between B6, BALB/c, and ART-2–/– T cells, supporting the notion that CD38 and not ART-2 is the principal NAD hydrolyzing cell surface enzyme (16). The finding that NAD is degraded during the time frame of our assays raises the possibility that NAD metabolites cause the effects ascribed to NAD.
FIGURE 3. Hydrolysis of NAD in the presence of T cells. B6, BALB/c, and ART-2–/– T cells were incubated in 200 μM [32P]NAD for the times indicated in complete RPMI 1640 without FBS. Supernatants were removed and analyzed by TLC. Plots show relative amounts of NAD and ADPR calculated from chromatographs on the left. The data shown are from one of three similar experiments.
To test whether ADPR or nicotinamide exert effects on T cells, both molecules were tested in T cell proliferation assays. Fig. 1C shows that proliferation of B6 and BALB/c T cells is inhibited by addition of 100 μM ADPR but not by nicotinamide. These results suggest that ADPR may be responsible for some of the effects seen with NAD, although 10-times higher concentrations of ADPR than of NAD are required to induce comparable effects.
Therefore, it was important to examine effects of ADPR on T cell survival. Fig. 4 shows that 1000 μM ADPR increases annexin V staining in the PI-negative B6 T cell population from 4.3 to 20.7% and PI staining from 44.7 to 78.6% at 24 h. This result is quite different from that seen with NAD in Fig. 2 in which there was a dramatic increase in annexin V staining but no increase in PI staining. In BALB/c T cells, ADPR fails to increase annexin V staining, and the PI staining cell population increases by only 12%. Again, this result contrasts to the dramatic effects of NAD in BALB/c T cells (Fig. 2). Therefore, it appears that while high concentrations of ADPR inhibit cell proliferation in both BALB/c and B6 T cells, only in B6 T cells does ADPR induce a significant increase in annexin V and PI staining. Thus, suppression of cell proliferation by ADPR does not correlate with cell death, and the effects of NAD on T cells cannot be solely explained by actions of its metabolite ADPR.
FIGURE 4. Effects of increasing doses of ADPR on B6 and BALB/c T cell viability. ADPR was added to T cells at the concentrations indicated. After 24 h cells were stained with annexin V, PI, and anti-CD3 and analyzed by FACS. Other conditions were as in Fig. 2. The data shown are from one of three repeat experiments.
NAD- but not ADPR-induced cell death is a rapid reaction
The finding that adding NAD and ADPR to B6 T cells is associated with different degrees of annexin V and PI staining at 24 h (Figs. 2 and 4) suggests that the two molecules promote cell death by different mechanisms. To examine this, the time course of cell death was examined in B6 T cells after the addition of either NAD or ADPR. Fig. 5 shows that 2 min after adding NAD, annexin V and PI staining reached a plateau. In contrast, attempts at demonstrating cell death by ADPR at these early time points were unsuccessful (data not shown). Therefore, NAD-induced cell death appears to be rapid while that induced by ADPR takes many hours. Results with BALB/c T cells confirm that NAD-induced cell death is exceedingly rapid as annexin V and PI staining plateau as early as 30 s after addition of NAD (Fig. 5). These results strongly suggest that NAD and ADPR induce cell death by different mechanisms and that B6 and BALB/c T cells differ in their sensitivity to these two molecules.
FIGURE 5. Time kinetics of NAD-induced cell death in B6 and BALB/c. NAD was added to T cells at a concentration of 500 μM. After the indicated times, the cells were stained with annexin V, PI, and anti-CD3 and analyzed by FACS. Other conditions were as in Fig. 2. The data shown are from one of three repeat experiments.
Features of apoptosis characterize slow but not rapid cell death induced by NAD
In light of the above results, it was of interest to identify mechanisms involved in inducing rapid and slow cell death. Cell death often proceeds by apoptotic mechanisms, and a hallmark of apoptosis is the activation of caspases. Therefore, we tested whether inhibition of caspases interferes with rapid or slow induction of cell death. B6 T cells were first incubated with caspase inhibitor Z-VADfmk and then different concentrations of NAD were added. Fig. 6A shows that after 24 h, compared with no addition, adding NAD increases the PI-negative, annexin V-positive cell population from 5.1 to 81.2%, and that Z-VADfmk reduces staining to 26.7%. Therefore, the mechanism by which NAD induces slow cell death appears to involve the action of caspases. Because slow cell death is also seen with ADPR, we assayed effects of Z-VADfmk on ADPR-induced cell death. Fig. 6A shows ADPR increases the PI-negative annexin V-positive population from 5.1 to 19.9%, which is suppressed by the inhibitor. ADPR also increases the PI staining population, i.e., from 45.6 to 78%, which is decreased to 52.2% by Z-VADfmk (Fig. 6A). These results suggest that the slow induction of cell death by both NAD and ADPR involves caspases.
FIGURE 6. Effect of Z-VADfmk on NAD- and ADPR-induced cell death. A, NAD and ADPR were added to B6 T cells at 500 μM and cells analyzed by FACS after 24 h. Where indicated cells were first incubated with 50 μM Z-VADfmk for 1 h before addition of NAD or ADPR. Cells were stained with annexin V, PI, and anti-CD3 and analyzed by FACS. Other conditions were as in Fig. 2. B, Effect of Z-VADfmk on NAD-induced cell death in B6 and BALB/c T cells in the 2- and 24-h assays. NAD was added to T cells at 500 μM, and the cells were analyzed after 2 or 24 h. Data are from PI-negative cells stained for CD3 and annexin V. Other conditions were as in A. All experiments were repeated three times, and one representative experiment is shown.
Therefore, we examined whether caspases are also involved in the rapid induction of cell death. Fig. 6B shows that annexin V staining in the PI-negative B6 T cell population increases from 4.9 to 13.7% within 2 h of NAD addition and that Z-VADfmk does not inhibit this effect. Therefore, annexin V staining induced during the first 2 h of incubation appears to be caspase independent. BALB/c T cells show entirely concordant results (Fig. 6B).
To further test the hypothesis that apoptosis is not involved in the rapid induction of death, arising from the addition of NAD, T cells were assayed for DNA fragmentation. Fig. 7 shows that 2 h after the addition of NAD or ADPR there is no increase in DNA fragmentation assayed by TUNEL staining in either B6 or BALB/c T cells. During this time period, annexin V staining increased from 3.5 to 19.7% in B6 cells and from 3.2 to 58.9% in BALB/c T cells. Therefore, cell death in the first 2 h after addition of NAD does not involve DNA fragmentation. In contrast, TUNEL staining increased significantly in BALB/c and B6 T cells 24 h after the addition of NAD (Fig. 7). Adding ADPR instead of NAD produced similar but less dramatic effects after 24 h (Fig. 7). Therefore, both NAD and ADPR induce slow cell death by an apoptotic pathway, associated with DNA fragmentation.
FIGURE 7. Induction of DNA fragmentation by NAD and ADPR in B6 and BALB/c cells. NAD or ADPR was added to T cells at 500 μM. After 24 h, the cells were stained with anti-CD3, annexin V, PI, and TUNEL (BrdU). Shown are results from CD3-gated, BrdU-stained cells. Annexin V staining is shown for the PI-negative, CD3-positive cell population. The data shown are from one of three repeat experiments.
ART-2 is required for rapid but not slow cell death induced by NAD
The finding that NAD but not ADPR induces rapid cell death is consistent with the notion that ADP-ribosylation is involved. However, because NAD also induces slow cell death, the question arises whether ADP-ribosylation contributes to the slow induction of cell death.
To find out, cell death was examined in T cells from wild-type NOD and NOD ART-2 gene-deleted mice. Fig. 8A shows that 2 h after adding NAD to NOD ART-2–/– T cells, there is no increase in annexin V staining shown in the dot plots or PI staining represented by percentages below the plots. In contrast, NAD effectively induces staining in T cells from wild-type NOD mice. Therefore, rapid cell death is not observed in T cells lacking ART-2, consistent with the requirement of ADP-ribosylation.
FIGURE 8. Effects of NAD and ADPR on T cells from wild-type NOD and ART-2–/– NOD mice. A, Effect of Z-VAD fmk on NAD- and ADPR-induced cell death in wild-type and ART-2–/– T cells in the 2- and 24-h assays. NAD or ADPR was added to T cells at 500 μM and analyzed after 2 or 24 h. Where indicated, cells were first incubated with 50 μM Z-VADfmk for 1 h before addition of NAD or ADPR. Cells were stained with annexin V, PI, and anti-CD3 and analyzed by FACS. Dot plots show data for PI-negative cells stained for CD3 and annexin V. Note that the percentages of PI-positive T cells from each analysis and condition are shown as percentages under each panel. B, Induction of DNA fragmentation by NAD and ADPR in ART2–/– T cells. NAD or ADPR were added to T cells at 500 μM. After 24 h, the cells were stained with anti-CD3, annexin V, PI, and TUNEL (BrdU). Shown are results from CD3-gated, BrdU-stained cells. Annexin V staining is shown for the PI-negative, CD3-positive cell population. The data shown are from one of three repeat experiments.
In contrast, 24 h after the addition of NAD to NOD ART-2–/– T cells there is an increase in annexin V staining from 5.7 to 45%, which is suppressed to 11.4% by Z-VADfmk, seen in the dot plots. Although ADPR causes only a small increase of annexin V staining (5.7 to 10.2%), there is an increase in the PI staining cell population from 37.6 to 61.8% as indicated by the percentages under the individual panels. Both effects are decreased by Z-VADfmk. Similar results are seen in T cells from NOD control mice, but note that NAD-induced effects are more pronounced in the wild type compared with the ART-2–/– T cells. Thus, there is a slow induction of cell death that does not require a functional ART-2 enzyme. To further test this conclusion, DNA fragmentation was assayed in ART-2–/– T cells. Fig. 8B shows that both NAD and ADPR induce a moderate increase in TUNEL staining at 24 h. However, the two molecules likely act by different mechanisms, because ADPR but not NAD induces a substantial increase in PI staining at the 24-h time point (Fig. 8A).
Purinergic receptor P2X7 is required for NAD to induce rapid but not slow cell death
The finding that NAD degradation product ADPR mimics some of the effects induced by NAD, points to the involvement of purinergic receptors. Among the family of purinergic receptors, P2X7 is widely expressed on cells of the hemopoietic cell lineage (17, 18, 19, 20). Recent data also suggests that P2X7, whose well-established ligand is ATP, expresses low affinity for ADP (21). Moreover, inhibitors of P2X7 receptors can block induction of cell death by NAD (10). Taken together, these observations make P2X7 receptors a prime candidate to mediate effects of NAD and ADPR. To examine whether the rapid or slow mechanism of cell death, induced by NAD or ADPR, involve the P2X7 receptor, use was made of the finding that KN-62 inhibits within minutes the signaling function of this receptor (15, 22). BALB/c T cells were incubated with KN-62 for 10 min, followed by addition of NAD; 15 min later, T cells were stained for annexin V and PI. Fig. 9A shows that KN-62 completely inhibits NAD-induced annexin V and PI staining in this short-term assay. Similar results were obtained with B6 T cells (data not shown). However, we were not able to test KN62 in the slow induction of death because of its toxicity (data not shown).
FIGURE 9. A, Effect of KN-62 on NAD-induced cell death in BALB/c T cells. NAD was added to T cells at 500 μM. After 15 min, the cells were stained and subjected to FACS analysis. Where indicated, cells were first incubated with 20 μM KN-62 for 10 min before addition of NAD. Cells were stained with annexin V, PI, and anti-CD3 and analyzed by FACS. Values in the quadrants represent percentages. Panels show data from PI-negative cells stained for CD3 and annexin V, as well as total cells analyzed for PI staining and forward light scatter (FSC). B and C, Induction of cell death by NAD and ADPR in T cells lacking P2X7 receptors. NAD or ADPR were added to T cells from B6 and P2X7–/– mice at 500 μM. After 2 and 24 h, cells were stained with CD3, annexin V, and PI and analyzed by FACS. Annexin V staining is shown for the PI-negative, CD3-positive cell population. The data shown are from one of three repeat experiments.
Therefore, we examined T cells from P2X7 receptor gene-deleted B6 mice (12). Fig. 9, B and C, shows that 2 h after addition of NAD, B6 (Fig. 9B) but not P2X7–/– (Fig. 9C) T cells displayed increased annexin V and PI staining. At 24 h, significant annexin V and PI staining is seen in B6 cells (69.6%) and somewhat lower staining is seen in P2X7–/– T cells (51%). This result shows that deleting expression of P2X7 receptors prevents NAD from rapidly inducing cell death. In contrast, NAD still induced the slow cell death in the P2X7–/– T cells. Results with ADPR are less striking. Twenty-four hours after adding ADPR, annexin V staining had increased from 8.2 to 15.2% for B6 cells, but was only 7.4% in P2X7–/– T cells. Thus, a functional P2X7 receptor is required for NAD to rapidly induce cell death but not slow death. In contrast, the P2X7 receptor may play a role in the mechanism by which ADPR induces slow cell death.
P2X7 receptor function is not impaired in etheno-NAD-treated cells
Two hypotheses arise to explain how cell surface ADP-ribosylation induced by NAD leads to rapid cell death. One is that the ADP-ribosylation of a protein modulates its function. The other is that ADP-ribosyl groups serve as ligands for a receptor. To test these hypotheses, we assessed effects of ADP-ribosylation on P2X7 receptor-triggered cell death by its ligand ATP. As previously reported (13), incubating T cells with etheno-NAD leads to etheno-ADP-ribosylation (Fig. 1, A and B) but not cell death. In fact, etheno-ADP-ribosylation renders T cells resistant to actions of NAD (10), presumably by filling the sites that are ADP-ribosylated upon the addition of NAD. Fig. 10 shows that while incubation of BALB/c T cells with etheno-NAD does not induce annexin V or PI staining, it does block the action of a subsequent addition of NAD. In contrast, etheno-NAD only slightly blocks rapid induction of annexin V by ATP and PI staining. These results are consistent with the notion that cell surface ADP-ribosyl groups act as ligands for P2X7 receptors, but not with the possibility that ADP-ribosylation acts by modifying protein function.
FIGURE 10. Effects of ATP on BALB/c T cells after incubation with etheno-NAD. Etheno-NAD was added to T cells at 300 μM. After 30 min, cells were washed and then incubated with ATP or NAD at the indicated concentrations. After 90 min, the cells were stained with annexin V, PI, and anti-CD3 and analyzed by FACS. Values in the quadrants represent percentages. Data from PI-negative cells stained for CD3 and annexin V are shown. The data is from one of three repeat experiments.
Discussion
The observation that murine T cells express cell surface ADP-ribosyltransferase activity and that by providing the enzyme with the substrate NAD leads to modification of cell surface proteins, which is associated with inhibition of T cell functions and induction of cell death (1, 2, 3, 4), has raised intriguing questions as to the physiological relevance of this reaction. An important issue is whether NAD concentrations that trigger cell death can be reached in vivo. Although intracellular NAD approaches 1000 μM, levels in blood are only 0.13 μM (23, 24). As shown here and elsewhere (1, 4), concentrations as low as 1 μM NAD can exert effects on T cells. Therefore, NAD levels one-thousandth of those inside cells and 10 times higher then those in circulation can mediate suppressive effects. Moreover, this report demonstrates that NAD need be present for only seconds for effects to be evident. Thus, effective NAD levels could well be achieved during tissue injury or by constitutive release from live cells.
Another question is whether induction of death arises from NAD itself or its breakdown products. The results above confirm the observation (11) that T cells substantially hydrolyze NAD to ADPR, and that upon prolonged incubation ADPR exerts effects similar but not identical to those induced by NAD. It should also be noted that ADPR is metabolized to adenosine (25), which could act on T cells via action on purinergic receptors. Indeed effects of adenosine on T cells via action on A2A receptors are well documented in the literature (for review see Ref.26). Therefore, effects induced by the addition of NAD, particularly in incubations lasting 24 h, may very well result from the combined effects of NAD and its catabolic derivatives.
This report demonstrates that NAD can induce annexin V and PI staining in T cells by different mechanisms that lead to either rapid or slow cell death. Rapid cell death is not inhibited by caspase inhibitor Z-VADfmk, and therefore, is reminiscent of necrosis. Slow cell death is associated with DNA fragmentation and is inhibited by caspase inhibitor Z-VADfmk, and therefore, has features of apoptosis. It should be noted that while both NAD and ADPR slowly induce apoptotic death, they do not appear to act by identical mechanisms. This again may be due to action of NAD metabolites on purinergic receptors such as A2A receptors or others yet to be identified (26)
It is interesting that adding NAD to rat T cells leads to autoADP-ribosylation of cell surface ART-2 but apparently no other proteins (27, 28). Although adding either NAD or ADPR to rat T cells inhibits their proliferation, NAD does not induce a rapid increase in annexin V staining (28). Therefore, it appears that autoADP-ribosylation of cell surface ART-2 is not sufficient to induce rapid T cell death.
An enigmatic question is how NAD inhibits T cell functions and induces cell death. Initially, it appeared that effects observed after adding NAD to T cells correlated with cell surface ADP-ribosylation (2, 4). Subsequent data showed this correlation is not a tight one. B6 T cells undergo substantially more ADP-ribosylation than BALB/c T cells, yet NAD is less effective in promoting rapid cell death in B6 cells. An important difference between BALB/c and B6 mice is the differential expression of the two ART-2 genes ART-2a and ART-2b, which could be responsible for this effect. BALB/c mice express both ART-2a and ART-2b, whereas B6 mice express ART-2b only, because of a mutation in ART-2a (4). This raises the possibility that ADP-ribosylation by the ART-2a coded enzyme is more efficient in cell death induction then that by the ART-2b enzyme. However, this does not appear to be the case because NZW mice expressing only ART-2a possess T cells even less sensitive to NAD than B6 mice T cells (4). Therefore, it was important to demonstrate that ADP-ribosylation is critical for the induction of cell death. We show that deletion of the ART-2 genes in NOD mice blocks the rapid induction of cell death by NAD, consistent with results generated previously in a different mouse strain (7).
We present experimental evidence that deleting expression of the P2X7 gene blocks rapid cell death induction by NAD. These data prompt a plausible explanation for the apparent lack of correlation between the extent of T cell ADP-ribosylation and induction of rapid cell death in T cells from B6 and BALB/c mice. BALB/c T cells express the wild-type P2X7 receptor, which mediates efficient and rapid cell death upon binding ATP. In contrast, the B6 P2X7 receptor has a mutation in the TNFR1 deathlike domain resulting in much less efficient induction of cell death (20). Therefore, it appears that reduced induction of cell death by NAD in B6 compared with BALB/c T cells arises from the mutation altering the function of the P2X7 receptor in B6 mice.
Fig. 11 shows a model for the role of P2X7 receptors in mediating the effects arising from addition of NAD to cultured T cells. Using NAD, ART-2 attaches ADP-ribosyl groups to a cell surface protein near the P2X7 receptor or to the receptor itself. The ligand-binding site of the receptor engages the ADP-ribosyl groups, rapidly inducing cell death by pore formation. The data presented show that the death signal is induced within 30 s of NAD contact. Given the evidence that induction of the death signal involves activation of P2X7 receptors, it is likely that the formation of membrane pores is preceded by calcium fluxes. Indeed, incubation of mouse T cells with NAD induces calcium fluxes (10). Similar results were recently reported with human monocytes in which both NAD and ADPR induce calcium fluxes, but it was not determined which receptors may be involved (29). But, as previously demonstrated, the P2X7 receptor is also capable of inducing a slower, caspase-dependent apoptotic cell death (30, 31). Attachment of etheno-ADP-ribosyl groups fails to trigger rapid death via P2X7 activation due to the structural specificity of the P2X7 binding site, and by filling potential sites for ADP-ribosylations, etheno-NAD blocks the action of subsequently added NAD. However, after etheno-ADP-ribosylation, P2X7 receptors remain functional as ATP can still induce rapid cell death. In addition, metabolites of NAD and ADPR may react with other receptors that lead to slow cell death. Exposure of phosphatidyl serine, and thus binding of annexin V, characterizes the induction of both rapid and slow cell death.
FIGURE 11. A model for ecto-NAD-induced effects on T cells. Using NAD as a substrate, ART-2 attaches ADP-ribosyl groups to cell surface protein(s) near P2X7 receptors. Binding of covalently attached ADP-ribosyl groups to P2X7 induces cell death by rapid pore formation and possibly slower caspase activation. Attachment of etheno-ADP-ribosyl groups blocks sites that otherwise would be ADP-ribosylated upon addition of NAD. Etheno-ADP-ribosyl groups do not trigger P2X7 signaling and do not interfere with ATP signaling. NAD also undergoes hydrolysis to ADPR, which itself or when further catabolized may provide ligands for other purinergic receptors, inducing slower caspase-dependent apoptosis. Both the rapid and slow pathways involve exposure of phosphatidyl serine.
It is important to note that for ATP to trigger P2X7 signaling, concentrations are required that are considerably higher than those needed for NAD (32). Moreover, adding NAD to cells leads to ADP-ribosylations within seconds, which persist after NAD is removed (5, 10). Given that intracellular concentrations of NAD and ATP are similar (23, 33), it becomes plausible that NAD released during trauma may be at least as important as ATP in regulating T cell functions via P2X7 receptors.
The question arises as to whether the model proposed above explains observations in earlier reports. Consistent with our model may be the observation that depleting rats or mice of T cells expressing ART-2 accelerates development of autoimmune diabetes and lupus erythematosus (34, 35). This could be explained by a mechanism in which NAD released from cells controls effector T cells. However, we have observed that deletion of the tandem ART-2 genes in the NOD mouse strain used in the above experiments (36) has no effect on diabetes progression (data not shown), even though lymphocytes from the parent NOD strain are particularly sensitive to signaling through P2X7 receptors (37). Taken together, these observations indicate the need to further study the role of ecto-NAD and its metabolites in the induction of cell death.
Acknowledgments
We thank Dr. R. Santella for the gift of Ab 1G4 and Dr. C. Gabel for generously providing the P2X7 gene-deleted mice.
Footnotes
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.
1 This work was supported by Public Health Service Grants AI 40038 and AI 43954.
2 Address correspondence and reprint requests to Dr. Gunther Dennert at the current address: University of Southern California/Norris Comprehensive Cancer Center, P.O. Box 33800, 1441 Eastlake Avenue, M/S 73, Los Angeles, CA 90033-0800. E-mail address: dennert@usc.edu
3 Abbreviations used in this paper: ART, ADP-ribosyltransferase; NWNA, nylon wool nonadherent; ADPR, ADP-ribose; PI, propidium iodide.
Received for publication September 21, 2004. Accepted for publication November 18, 2004.
References
Wang, J., E. Nemoto, A. Y. Kots, H. R. Kaslow, G. Dennert. 1994. Regulation of cytotoxic T cells by ecto-nicotinamide adenine dinucleotide (NAD) correlates with cell surface GPI-anchored/arginine ADP-ribosyltransferase. J. Immunol. 153:4048.
Liu, Z. X., O. Azhipa, S. Okamoto, S. Govindarajan, G. Dennert. 2001. Extracellular nicotinamide adenine dinucleotide induces T cell apoptosis in vivo and in vitro. J. Immunol. 167:4942.
Han, M. K., Y. S. Cho, Y. S. Kim, C. Y. Yim, U. H. Kim. 2000. Interaction of two classes of ADP-ribose transfer reactions in immune signaling. J. Biol. Chem. 275:20799.
Adriouch, S., W. Ohlrogge, F. Haag, F. Koch-Nolte, M. Seman. 2001. Rapid induction of naive T cell apoptosis by ecto-nicotinamide adenine dinucleotide: requirement for mono(ADP-ribosyl)transferase 2 and a downstream effector. J. Immunol. 167:196.
Nemoto, E., Y. Yu, G. Dennert. 1996. Cell surface ADP-ribosyltransferase regulates lymphocyte function-associated molecule-1 (LFA-1) function in T cells. J. Immunol. 157:3341.
Kahl, S., M. Nissen, R. Girisch, T. Duffy, E. H. Leiter, F. Haag, F. Koch-Nolte. 2000. Metalloprotease-mediated shedding of enzymatically active mouse ecto-ADP-ribosyltransferase ART2.2 upon T cell activation. J. Immunol. 165:4463.
Ohlrogge, W., F. Haag, J. Lohler, M. Seman, D. R. Littman, N. Killeen, F. Koch-Nolte. 2002. Generation and characterization of ecto-ADP-ribosyltransferase ART2.1/ART2.2-deficient mice. Mol. Cell. Biol. 22:7535.
Liu, Z. X., Y. Yu, G. Dennert. 1999. A cell surface ADP-ribosyltransferase modulates T cell receptor association and signaling. J. Biol. Chem. 274:17399.
Okamoto, S., O. Azhipa, Y. Yu, E. Russo, G. Dennert. 1998. Expression of ADP-ribosyltransferase on normal T lymphocytes and effects of nicotinamide adenine dinucleotide on their function. J. Immunol. 160:4190
Seman, M., S. Adriouch, F. Scheuplein, C. Krebs, D. Freese, G. Glowacki, P. Deterre, F. Haag, F. Koch-Nolte. 2003. NAD-induced T cell death: ADP-ribosylation of cell surface proteins by ART2 activates the cytolytic P2X7 purinoceptor. Immunity 19:571.
Bortell, R., J. Moss, R. C. McKenna, M. R. Rigby, D. Niedzwiecki, L. A. Stevens, W. A. Patton, J. P. Mordes, D. L. Greiner, A. A. Rossini. 2001. Nicotinamide adenine dinucleotide (NAD) and its metabolites inhibit T lymphocyte proliferation: role of cell surface NAD glycohydrolase and pyrophosphatase activities. J. Immunol. 167:2049.
Solle, M., J. Labasi, D. G. Perregaux, E. Stam, N. Petrushova, B. H. Koller, R. J. Griffiths, C. A. Gabel. 2001. Altered cytokine production in mice lacking P2X(7) receptors. J. Biol. Chem. 276:125.
Krebs, C., W. Koestner, M. Nissen, V. Welge, I. Parusel, F. Malavasi, E. H. Leiter, R. M. Santella, F. Haag, F. Koch-Nolte. 2003. Flow cytometric and immunoblot assays for cell surface ADP-ribosylation using a monoclonal antibody specific for ethenoadenosine. Anal. Biochem. 314:108.
Matsuoka, S., H. Tsurui, M. Abe, K. Terashima, K. Nakamura, Y. Hamano, M. Ohtsuji, N. Honma, I. Serizawa, Y. Ishii, et al 2003. A monoclonal antibody to the 2 domain of murine major histocompatibility complex class I that specifically kills activated lymphocytes and blocks liver damage in the concanavalin A hepatitis model. J. Exp. Med. 198:497.
Humphreys, B. D., C. Virginio, A. Surprenant, J. Rice, G. R. Dubyak. 1998. Isoquinolines as antagonists of the P2X7 nucleotide receptor: high selectivity for the human versus rat receptor homologues. Mol. Pharmacol. 54:22.
Howard, M., J. C. Grimaldi, J. F. Bazan, F. E. Lund, L. Santos-Argumedo, R. M. Parkhouse, T. F. Walseth, H. C. Lee. 1993. Formation and hydrolysis of cyclic ADP-ribose catalyzed by lymphocyte antigen CD38. Science 262:1056.
Mutini, C., S. Falzoni, D. Ferrari, P. Chiozzi, A. Morelli, O. R. Baricordi, G. Collo, P. Ricciardi-Castagnoli, F. Di Virgilio. 1999. Mouse dendritic cells express the P2X7 purinergic receptor: characterization and possible participation in antigen presentation. J. Immunol. 163:1958.
Collo, G., S. Neidhart, E. Kawashima, M. Kosco-Vilbois, R. A. North, G. Buell. 1997. Tissue distribution of the P2X7 receptor. Neuropharmacology 36:1277
Mehta, V. B., J. Hart, M. D. Wewers. 2001. ATP-stimulated release of interleukin (IL)-1 and IL-18 requires priming by lipopolysaccharide and is independent of caspase-1 cleavage. J. Biol. Chem. 276:3820.
Adriouch, S., C. Dox, V. Welge, M. Seman, F. Koch-Nolte, F. Haag. 2002. Cutting edge: a natural P451L mutation in the cytoplasmic domain impairs the function of the mouse P2X7 receptor. J. Immunol. 169:4108.
Chakfe, Y., R. Seguin, J. P. Antel, C. Morissette, D. Malo, D. Henderson, P. Seguela. 2002. ADP and AMP induce interleukin-1 release from microglial cells through activation of ATP-primed P2X7 receptor channels. J. Neurosci. 22:3061.
Gargett, C. E., J. S. Wiley. 1997. The isoquinoline derivative KN-62 a potent antagonist of the P2Z-receptor of human lymphocytes. Br. J. Pharmacol. 120:1483.
Jacobson, E. L., M. K. Jacobson. 1997. Tissue NAD as a biochemical measure of niacin status in humans. Methods Enzymol. 280:221.
Kim, U. H., M. K. Kim, J. S. Kim, M. K. Han, B. H. Park, H. R. Kim. 1993. Purification and characterization of NAD glycohydrolase from rabbit erythrocytes. Arch. Biochem. Biophys. 305:147.(Hiroki Kawamura, Fred Asw)