Microchemical Core Facility, Department of Biology, San Diego State University, 5500 Campanile Drive, San Diego, CA 92182-1030, U.S.A. E-mail: email@example.com
Extracellular ATP is a potent signaling molecule that modulates a myriad of cellular functions through the activation of P2 purinergic receptors and is cytotoxic to a variety of cells at higher concentrations. The mechanism of ATP-elicited cytotoxicity is not fully understood. In this study, we investigated the effect of extracellular ATP on the human hepatoma Li-7A cells.
We observed a time- and dose-dependent growth inhibition of Li-7A cells by ATP, which is accompanied by an increase in the active form of caspase-3 as well as increased cleavage of its substrate, poly (ADP-ribose) polymerase. The cytotoxic effect of extracellular ATP was not mediated by the P2X7 receptor, since (1) the effect was not abolished by the P2X7 receptor antagonists oxidized ATP and KN-62, and (2) extracellular ADP, AMP, and adenosine were also cytotoxic.
We found that ATP and ADP were degraded to adenosine by Li-7A cells and that treatment of Li-7A cells by adenosine resulted in growth inhibition and caspase-3 activation, indicating that adenosine is the apoptotic agent. Using adenosine receptor agonists and antagonists, as well as inhibitors of adenosine transport and deamination, we showed that the cytotoxic effect of adenosine is specifically mediated by the A3 receptor even though transcripts of A1, A2A, A2B, and a splice variant of the P2X7 receptors were detected in Li-7A cells by RT–PCR.
Cytotoxicity caused by exogenous ATP and adenosine was completely abolished by the caspase-3 inhibitor Z-DEVD-FMK, demonstrating the central role of caspase-3 in apoptosis of Li-7A cells.
Extracellular adenosine and ATP elicit a variety of physiological responses in different cells. These effects are mediated by specific purinergic receptors for extracellular nucleosides or nucleotides. It is increasingly clear that purinergic signaling plays important roles in various pathological conditions such as cancer, ischemia-associated injury, inflammation, wound healing, drug toxicity, and pain. For the development of therapeutic agents, it is important to define the involvement of specific receptor subtypes in different tissues.
The effect of extracellular purines on transformed cells was first reported in the 1970s. The observation that nucleosides and phosphate esters traffic through the membranes of ATP-treated cells suggested that relatively large pores are formed in the cell membranes (Rozengurt et al., 1977; Heppel et al., 1985). Subsequently, it was shown that ATP-induced ion fluxes and membrane permeabilization also occur in normal cells, especially in cells of the immune system (Cockcroft & Gomperts, 1980; Steinberg & Silverstein, 1987; Steinberg & Di Virgilio, 1991). Interestingly, the effects of extracellular ATP on cytolytic T lymphocytes bear all the hallmarks of apoptosis, that is, surface blebbing and nuclear DNA fragmentation, in addition to dye uptake (Zheng et al., 1991), thus identifying extracellular ATP as an apoptotic agent.
Owing to the ubiquitous presence of a variety of ectonucleotidases (see Zimmermann (2001) for review), extracellular ATP can be degraded to adenosine, a ligand of the adenosine receptors. Adenosine has also been implicated in the induction of apoptosis in several cell types (for a review, see Chow et al., 1997). Extracellular adenosine acts mainly through the activation of the adenosine receptors, of which four subtypes have been cloned from various tissues and species (reviewed in Ralevic & Burnstock, 1998). These receptors, all of which are G protein-coupled, are classified into four categories, A1, A2A, A2B, and A3, on the basis of their molecular structures, their distinct pharmacological profiles, and tissue distributions. Agonists selective for these receptors are known. Among these, only the A3 receptor agonists, such as IB-MECA and Cl-IB-MECA, have been shown to induce apoptosis in leukemia cells and cardiac myocytes (Kohno et al., 1996a,1996b; Shneyvays et al., 2000).
In the present study, we have investigated the effect of extracellular nucleotides in a human hepatoma cell line, Li-7A, in an attempt to understand the mechanism of extracellular nucleotide-induced cell death of cancer cells. Our results suggest that extracellular ATP induces apoptosis of Li-7A cells after being degraded to adenosine. We further show that adenosine-induced apoptosis in Li-7A cells is an A3 receptor-dependent process and that activation of the caspase-3 cascade is required.
Human hepatoma Li-7A cells (Knowles, 1988a) were grown in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal calf serum and antibiotics (50 U ml−1 penicillin, 50 μg ml−1 streptomycin). Human embryonic kidney cells (HEK293) stably transfected with the human P2X7 receptor (HEK-P2X7 cells) (kindly provided by Dr G.R. Dubyak, Case Western Reserve University) were grown in DMEM containing 5% new born calf serum and 5% fetal calf serum, streptomycin, penicillin, and 25 μg ml−1 hygromycin. Cells were grown in 24-well or 10-cm culture plates, in a humidified atmosphere of 5% CO2 at 37°C.
The effect of nucleotides and nucleosides on the proliferation of Li-7A cells was evaluated by the MTT bioassay (Cory et al., 1991). After Li-7A cells were grown to confluency in 24-well plates, the cells were washed twice with serum-free DMEM and then cultured in the absence or presence of the ligands of the purinergic receptors. At the indicated times, MTT tetrazolium salt with phenazine methosulfate was added directly to the culture media and allowed to incubate with the cells for 2–3 h. Mitochondrial dehydrogenases of viable cells will convert MTT into a color-dense formazan. The media from each well were collected in microfuge tubes and detached cells were recovered by centrifugation. DMSO was added to the cell pellets in the tubes as well as attached cells remaining in the wells to dissolve the insoluble formazan. The DMSO solutions were combined and the absorbance was measured at 550 nm with a reference wavelength of 630 nm after appropriate dilution. At least three independent experiments were performed. Data presented were selected from representative experiments.
Tdt-mediated dUTP nick end labeling (TUNEL) assay
TUNEL staining was performed using an in situ cell detection kit, BrdUTP-FragEL™, according to the manufacturer's instruction (Oncogene Research Products, San Diego, CA, U.S.A.). The cells were plated on Lab-Tek chamber slides, grown overnight and then treated with or without ATP. Cells were washed, fixed with 4% paraformaldehyde in PBS solution, and incubated with hydrogen peroxide to block endogenous peroxidase. After rinsing, the cells were incubated with the TUNEL reaction mixture. Thereafter, the cells were washed and incubated with biotinylated monoclonal anti-BrdU and a streptavidin-horseradish peroxidase conjugate, followed by incubation with diaminobenzidine (DAB)-substrate solution, and then counterstained with methyl green.
cDNA microarray analysis
Total cellular RNA was prepared from Li-7A cells cultured in the absence or presence of 6 mM ATP for 24 h. A measure of 10 μg of the total RNA was used as template for [32P] cDNA probe synthesis. The RNA was first annealed with a primer mix specific for apoptosis-related genes and the cDNA probes were synthesized according to the protocols provided by the manufacturer. The cDNA probes were denatured and hybridized with the apoptosis-related cDNA microarrays (Super Array Inc., Bethesda, MD, U.S.A.). Hybridization was carried out at 68°C overnight. The membrane was washed twice with wash solution (2 × SSC, 1% SDS) for 20 min at 68°C. The membrane was wrapped in a plastic wrap and exposed to an X-ray film with an intensifying screen at −70°C, developed or scanned by a phosphorimager. Each array is composed of 23 marker genes in duplicates, with two positive controls, β-actin and (glyceraldehyde-3-phosphate dehydrogenase (GAPDH)), and a negative control, bacterial plasmid pUC18. In addition, it contained cDNA gene fragments of the Bcl-2 gene family, caspases, caspase-regulatory molecules, as well as Fas, E2F, TRAIL, p21, pig, p53, and NF-κB. Relative amounts of mRNA transcripts were quantified using a Computing Densitometer™ scanner and ImageQuant™ V3.3 (Molecular Dynamics), and compared with the signals derived from GAPDH and β-actin.
Cells were lysed and protein extraction was performed. Protein concentrations of the cell lysates were determined by the Dc protein assay (Bio-Rad, Hercules, CA, U.S.A.). Cell proteins (25 μg) were separated in 10 or 13% sodium dodecyl sulfate–polyacrylamide gel, and electrophoretically transferred to a PVDF membrane (NEN Life Science, Boston, MA, U.S.A.). The membranes were blocked with 5% nonfat milk, washed, and subsequently incubated with a mouse monoclonal antibody against poly (ADP-ribose) polymerase (PARP) (Santa Cruz Biotech, Santa Cruz, CA, U.S.A.) at 1 : 1000 dilution, or a rabbit polyclonal antibody against caspase-3 (Santa Cruz Biotech, Santa Cruz, CA, U.S.A.) at 1 : 1000 dilution. After washing, the membrane was incubated with horseradish peroxidase-conjugated sheep anti-mouse antibody or donkey anti-rabbit antibody (Amersham Pharmacia Biotech, Piscataway, NJ, U.S.A.), and then visualized by enhanced chemi-luminescence (ECL) according to the manufacturer's recommendations (Amersham Pharmacia Biotech, Piscataway, NJ, U.S.A.).
HPLC separation of adenosine and adenine nucleotides
Nucleotides were separated by HPLC (Waters, ‘Millennium’ system) on a DNApac™ PA-100 anion exchange column (4 × 250 mm2, Dionex), using a 45-min linear gradient deve-loped from 20 mM sodium phosphate (pH 7.0). to 20 mM sodium phosphate with 1 M sodium chloride (pH 7.0). The method allowed clear separation of ATP, ADP, AMP, and adenosine. Peaks in the eluates were identified by comparison to known standards for their characteristic retention times.
The presence of the transcripts of human P2X7, Al, A2A, A2B, and A3 receptors in Li-7A cells were confirmed using RT–PCR. Total cellular RNA from Li-7A and HEK-P2X7 cells was prepared using the RNAzol method (Teltest, Firendwood, TX, U.S.A.), according to the manufacturer's instruction. First-strand cDNA was synthesized from 1 μg of total cellular RNA, using an oligo(dT)18 primer and Superscript II reverse transcriptase (Life Technologies, Gaithersburg, MD, U.S.A.) at 42°C for 1 h. PCRs were performed in a 50 μl reaction mixture containing 0.5 μl Taq DNA polymerase (5 U μl−1), 1 μl cDNA from the cDNA synthesis reaction mixture, 0.2 mM dNTP, and 50 nM of various primer pairs specific for the human P2X7 receptor (accession number Y09561), and the human Al, A2A, A2B (accession number BC025722), and A3 receptors. For the human P2X7 receptors, two forward primers directed to bases l–23 (5′-GCAGGGAGGGAGGCTGTCACCAT-3′), 85–106 (5′-CCAGAGCATGAATTATGGCACC-3′) and two reverse primers directed to bases 1321–1440 (5′-TATCCCTGGATCTAGGAGTC-3′), 1860–1885 (5′-GCCTGGCTTCAGTAAGGACTCTTGAA-3′) were used. Forward primer 5′-CAACATTGGGCCACAGACCT-3′ and reverse primer 5′-TAGGTAAGGATGCTGGGCTT-3′ (Suzuki et al., 1998) were used for the Al receptor, giving a 606 bp product. Forward primer 5′-AGATGGAGAGCCAGCCTCTGC-3′, and reverse primer 5′-GCTAAGGAGCTCCACGTCTGG-3′ (Suzuki et al., 1998) were used for the A2A receptor, giving a 535 bp product. Forward primer 5′-CAAGTCACTGGCCATGA-TTG-3′ (785–805) and reverse primer 5′-CCACCA-TAAACAAGGCAGAC-3′ (1362–1381) were designed and used for the A2B receptor, giving a 596 bp product. Forward primer 5′-ACCCCCATGTTTGGCTG-3′ and reverse primer 5′-GCACAAGCTGTGGTACCTCA-3′ (Kohno et al., 1996a) were used for the A3 receptor, giving a 361 bp product. The PCR products were analyzed after electrophoresis on 1.5% agarose gel and staining by ethidium bromide. The PCR products were also cloned into the pcDNA3.1 vector (TA cloning kit, Invitrogen, Carlsbad, CA, U.S.A.), according to the manufacturer's instructions, and sequenced completely using an automated DNA sequencer.
Statistical comparison was carried out by the Student's t-test. Data are expressed as means±s.d. P-values less than 0.05 were considered as significant.
DMEM, fetal and new born calf sera were purchased from GIBCO Life Technologies (Gaithersburg, MD, U.S.A.). All nucleotides and nucleosides, dipyridamole, N6-cyclohexyladenosine (CHA), 5-(N-cyclopropyl) carboxamidoadenosine (CPCA), 8-cyclopentyl-l,3-dipropylxanthine (DPCPX), alloxazine, l-(N, O-bis[5-isoquinolinesulfonyl]-N-methyl-L-tyrosyl)-4-phenylpiperazine (KN-62), and periodate-oxidized ATP (oxATP) were obtained from Sigma (St Louis, MO, U.S.A.). Adenosine receptor agonists and antagonists, 2-chloro-N6-(3-iodobenzyl)-adenosine-5′-N-methylurinamide(2-C1-IB-MECA), ZM241385, MRS 1220, and MRS 1523 were from Tocris (Ellisville, MO, U.S.A.). Protease inhibitors, Z-DEVD.fmk, Z-FA.fmk, and N-acetyl-leucyl-leucyl-norleucinal (ALLN), and adenosine deaminase inhibitors, coformycin and erythro-9-(2-hydroxy-3-nonyl)adenine, HC1(EHNA), were from Calbiochem (La Jolla, CA, U.S.A.).
Extracellular ATP induces cytotoxicity in Li-7A cells
To investigate if ATP exerts cytotoxic effect on Li-7A cells, we treated Li-7A cells with 0–10 mM ATP at various time intervals, and used the MTT colorimetric assay to quantify the effect of ATP on Li-7A cell proliferation. As seen in Figure 1, ATP caused a time- and dose-dependent inhibition of the proliferation of Li-7A cells. Treatment of Li-7A cells by ATP for only 7 h had no effect even at 10 mM ATP. Cell growth inhibition was observed when exposure to extracellular ATP was prolonged. The effective ATP concentration for 50% inhibition (EC50) of Li-7A cell growth after 39 h is 5 mM. In the following experiments, 6 mM (higher than EC50) of ATP was used as the standard concentration.
We then verified that ATP-induced cell death in Li-7A cells was the result of apoptosis. Apoptosis is characterized by internucleosomal degradation of genomic DNA. DNA cleavage and free 3′-OH groups generated by cellular endo-nucleases in apoptotic cells can be determined using in situ TUNEL assay (Gavriell et al., 1992). A dark brown DAB signal indicates positive staining, while shades of blue-green to greenish tan indicate a nonreactive cell. Figure 2 shows the result of a representative experiment demonstrating that the number of TUNEL-positive cells increased with increasing ATP concentrations. These results were confirmed by FACS analysis (data not shown), where cells were stained with propidium iodide and FITC-conjugated annexin V antibodies, which only bind to apoptotic cells. We conclude that extracellular ATP induces apoptosis in Li-7A cells.
Extracellular ATP induces caspase-3 gene expression
To define the molecular basis of the cytotoxic action of ATP, we examined the effect of ATP on the expression of apoptosis-related genes using cDNA microarray analysis. Through side-by-side hybridization with cDNA probes prepared from RNAs of untreated and ATP-treated Li-7A cells, the expression profiles of apoptosis-related genes were compared. The results in Figure 3 show that the bacterial plasmid pUC18 was not expressed in either control or ATP-treated Li-7A cells (spots 1G and 2G) and served as negative control. ATP treatment of Li-7A caused no changes in the expression of the housekeeping genes, β-actin (spots 3G and 4G) and GAPDH (spots 5G, 6G, 7G, 8G, 8E, and 8F). Importantly, ATP treatment caused a significant upregulation of caspase-3 gene expression (spots 3C and 3D) in Li-7A cells. The expression of other cDNAs showed either slight or no changes after ATP treatment.
Activation of caspase cascade in ATP-induced cell death of Li-7A cells
Caspases have been implicated as the principal mediators of apoptosis. Increased caspase 3/CPP32 activity results in increased cleavage of its substrate PARP (He et al., 1998). Caspase-3, like the other caspases, is synthesized as an inactive zymogen. Upon activation, it is converted to a tetrameric complex composed of two heterodimeric subunits of 17 kDa. Using Western blot analysis, we demonstrated the appearance of the 17 kDa caspase-3 subunits in Li-7A cells after 39 h of treatment with 6 mM ATP (Figure 4a). As observed for other cells undergoing apoptosis, significant amount of the 32 kDa zymogen remained (He et al., 1998; Wen et al., 2003). Nevertheless, the active caspase-3 cleaved its substrate, the 117 kDa PARP, into its characteristic 89 kDa fragment (Figure 4a). Figure 4b shows that the 17 kDa form of caspase-3 as well as the 89 kDa cleavage product of PARP were detected in Li-7A cells after being treated with 2–8 mM ATP for 36 h.
Effect of other nucleotides and adenosine on Li-7A cell proliferation
To determine if the cytotoxic effect of ATP on Li-7A cells is mediated by the P2X7 receptor, we determined the effects of two P2X7 receptor antagonists, oxATP (Murgia et al., 1993) and KN62 (Humphreys et al., 1998), on ATP-induced cell death in Li-7A cells. In previous studies using HEK cells stably transfected with the human P2X7 receptor, we showed that ATP-induced apoptosis was blocked by 1 mM oxATP and 3.5 μM KN62 (Wen et al., 2003). We found that oxATP itself inhibited the growth of Li-7A cells at 0.25 and 1 mM. Inhibition of growth by oxATP did not involve the activation of caspase-3. ATP-induced caspase-3 activation was also not diminished by oxATP (data not shown). Furthermore, the more potent P2X7 receptor antagonist, KN-62, was ineffective in blocking ATP-induced cell death of Li-7A cells (data not shown). These results indicate that ATP-induced cell death in Li-7A cells is not mediated by the P2X7 receptor.
We then investigated the effects of other purinergic receptor ligands on Li-7A cell proliferation. Figure 5a shows that 6 mM adenosine, AMP, and ADP inhibited cell growth to the same extent as ATP, whereas inhibition by 6 mM UTP was less. Similar to the effect of ATP, cytotoxicity induced by ADP, AMP, and UTP also involved activation of caspase-3, as indicated by the appearance of the 17 kDa caspase subunit and the 89 kDa PARP cleavage product (Figure 5b).
Effect of adenosine on Li-7A cell proliferation
The observation that extracellular AMP, which is not a ligand of P2 receptor, was effective in inducing the cell death of Li-7A cells, suggests that adenosine, which can be produced from AMP by the action of 5′-nucleotidase, is most likely the apoptotic agent. The effects of ATP and ADP could also be explained by the generation of adenosine from these nucleotides by the action of ectonucleotidases in these cells (Knowles, 1988b). Figure 6 shows that adenosine was indeed produced from ATP and ADP added to Li-7A cells. HPLC analysis of adenine metabolites in the culture media showed that ∼50% of the added ATP and all of the added ADP were converted to adenosine after 24 h.
Figure 7a shows that adenosine was as effective as ATP in inducing cell cytotoxicity in Li-7A cells. While treatment for 24 h by 4 mM ATP caused an ∼30% inhibition of cell growth (Figure 1), 4 mM adenosine inhibited cell growth by ∼70% (Figure 7a). Similar to the effect of ATP, treatment of Li-7A cells by 6 and 8 mM adenosine for 15 h also resulted in the cleavage of PARP into the characteristic 89 kDa fragment and the conversion of the inactive caspase-3 precursor to the 17 kDa subunit (Figure 7b). The disappearance of the 32 kDa caspase zymogen was more pronounced in adenosine-treated than in ATP-treated Li-7A cells. These results support the conclusion that it is the metabolite of ATP, adenosine, rather than ATP per se that causes the cell death of Li-7A cells.
To further establish that adenosine-induced cell death in Li-7A cells is caspase-3-dependent, the effect of the caspase-3 inhibitor Z-DEVD.fmk was tested. Figure 8 shows that 200 μM Z-DEVD.fmk effectively inhibited both ATP- and adenosine-induced cell death by 84%. On the other hand, 200 μM Z-FA.fmk, a structural analog of Z-DEVD.fmk in which the amino-acid sequence DEVD has been replaced, and ALLN, a proteasome inhibitor, were ineffective. Taken together, these data are consistent with the conclusion that caspase-3 activation is required for adenosine-mediated apoptosis of Li-7A cells.
Effect of adenosine deaminase and nucleoside transport inhibitors on ATP- and adenosine-induced cell death
To demonstrate that apoptosis of Li-7A cells is induced specifically by extracellular adenosine, we investigated the effects of inhibitors of adenosine deaminase and adenosine uptake on adenosine-induced cytotoxicity. Li-7A cells were incubated with ATP or adenosine in the absence or presence of adenosine deaminase inhibitors EHNA and coformycin. Figure 9 shows that 300 μM EHNA had no effect on cytotoxicity induced by either ATP or adenosine, whereas equal concentration of coformycin, a more potent adenosine deaminase inhibitor, significantly enhanced the cytotoxic effects of ATP and adenosine. This observation indicates that the cytotoxic effects of ATP and adenosine on Li-7A cells were mediated by adenosine, and not its deaminated metabolite inosine.
It has been shown that adenosine can exert its cytotoxic effect intracellularly after it is transported into cells (Chow et al., 1997; Schrier et al, 2001; 2002). If this is the case, then adenosine-induced cytotoxicity in Li-7A cells will be abolished by the nucleoside transport inhibitor dipyridamole, or another substrate of the transporter, uridine. We found that neither dipyridamole nor uridine had any effect on adenosine-induced cytotoxicity in Li-7A cells (data not shown), supporting the conclusion that the cytotoxic effect is exerted by extracellular adenosine.
Effect of adenosine receptor agonists and antagonists on ATP- and adenosine-induced cell death of Li-7A cells
To determine whether the apoptotic effect of adenosine was mediated by a specific adenosine receptor, we tested several adenosine receptor agonists. N6-cyclohexyladenosine (CHA), CPCA, and 2-C1-IB-MECA are selective agonists for the Al, A2, and A3 receptors, respectively. At concentrations of 300 μM, CHA and CPCA did not induce cell death. However, in the concentration range of 100–300 μM, the adenosine A3 receptor agonist 2-C1-IB-MECA induced cell death (Figure 10a). To further confirm that adenosine-induced cell death in Li-7A cells is mediated by the A3 receptor, various adenosine receptor antagonists were tested. DPCPX, ZM241385, alloxazine, and MRS1220 are selective antagonists for the Al, A2A, A2B, and A3 receptors, respectively. Figure 10b shows that DPCPX, ZM241385, and alloxazine did not block adenosine-induced cell death at 200 μM. Only MRS 1220, a human A3 receptor-selective antagonist, partially inhibited cell death induced by either adenosine or the A3 receptor agonist Cl-IB-MECA. We conclude that the cytotoxic effect of adenosine on Li-7A cells is mediated by the A3 receptor.
RT–PCR of human adenosine receptors and P2X7 receptor of Li-7A cells
The presence of the A3 receptor in Li-7A cells can also be demonstrated by RT–PCR. However, transcripts of the Al, A2A, and A2B receptors are present as well (Figure 11a). Interestingly, we also detected the presence of the P2X7 receptor transcript in Li-7A cells by RT–PCR. A 1.8 kb PCR product was obtained using the forward primer directed to bases 85–106 and reverse primer to bases 1860–1885 (Figure 11b, lane 4), and a 1.35 kb product was obtained using forward primer directed to bases 85–106 and reverse primers directed to bases 1321–1440 (Figure 11b, lane 3). The amplified partial cDNAs were subcloned and their sequences were found to be identical to that reported for the same region of the human P2X7 receptor (Rassendren et al., 1997). On the other hand, we failed to obtain a product corresponding to the full-length cDNA in RT–PCR reaction using the 5′-terminal forward primer directed to bases 1–23, designed from the same sequence (Figure 9b, lane 2), while an ∼1.9 kb PCR product could be obtained using cDNA prepared from HEK cells stably transfected with P2X7 cDNA (HEK-P2X7 cells) (Figure 9b, lane 1). These results suggest that a splice variant of the P2X7 receptor is present in Li-7A cells. Its functionality remains to be determined.
Based on our current understanding of signal transduction of extracellular ATP, plasma membrane permeability changes are mediated by the P2X receptors, which are ligand-gated ion channels. Two subtypes of the P2X receptors, P2X7 and P2X1, have been implicated in apoptosis (Chvatchko et al., 1996; Chow et al., 1997; Wen et al., 2003). P2X7 receptor is unique among the P2X receptors in that its activation results in large pores in the membrane that permit the fluxes of molecules of up to 900 Da. This receptor is most certainly involved in dye uptake of ATP-treated cells (Surprenant et al., 1996; Rassendren et al., 1997; Kim et al., 2001). Recently, we demonstrated the involvement of the P2X7 receptor in ATP-induced apoptosis in HEK293 cells stably transfected with the human P2X7 receptor cDNA (Wen et al., 2003). We showed that the apoptotic response of P2X7-HEK cells was only induced by ATP, while ADP and AMP were ineffective. We further showed that ATP-induced cell death could be prevented by two P2X7 receptor antagonists, that is, oxATP and KN62. In contrast to the HEK-P2X7 cells, the present study showed that cell growth inhibition of Li-7A cells was induced not only by extracellular ATP, but also extracellular ADP and AMP (Figure 5). Furthermore, the cytotoxic effect of extracellular ATP was not diminished by the P2X7 receptor antagonists. Since extracellular adenine nucleotides could be degraded to adenosine by the combined actions of several ectonucleo-tidases, these results indicated that adenosine might be the cytotoxic agent of Li-7A cells. HPLC analysis revealed that indeed the majority of the ATP and ADP added to the media were degraded to adenosine by the cells (Figure 6). The time required for adenosine production from ATP hydrolysis may explain the length of time (24–39 h) needed for the manifestation of apoptosis.
Adenosine has been shown to inhibit the growth of several types of cells (Chow et al., 1997; Schrier et al., 2001; 2002). There are two different underlying mechanisms of the growth-inhibitory actions of adenosine. Adenosine can exert its effect extracellularly, which is mediated by the adenosine receptors, as shown for HL-60 leukemia cells (Kohno et al., 1996a). Alternatively, adenosine causes cell growth inhibition after being transported into the cells, bringing about pyrimidine starvation (Lee & Jarvis, 1988; Pajor & Wright, 1992). To distinguish between these two possibilities, we conducted experiments in the presence of the nucleoside transporter inhibitor dipyridamole. Dipyridamole is expected to abolish the cytotoxic effect of adenosine on Li-7A cells if the nucleoside affects intracellular nucleotide metabolism. We found that dipyridamole is ineffective in blocking growth inhibition by adenosine, indicating that its effect is mediated by one of the adenosine receptors. This conclusion was supported by the finding that the A3 adenosine receptor agonist Cl-IB-MECA was effective in inhibiting Li-7A cell growth at concentrations (100 μM) significantly lower than that required for adenosine or the adenine nucleotides (2–6 mM) (Figure 10a, b). Furthermore, the cytotoxic effect of Cl-IB-MECA was partially blocked by the specific A3 receptor antagonist MRS 1220 (Figure 10b).
It is interesting that the A3 receptor is specifically involved in the apoptotic response of Li-7A cells to adenosine or ATP, since RT–PCR demonstrated the presence of all the four subtypes of the adenosine receptor, as well as a splice variant of the P2X7 receptor in these cells. The A3 receptor has also been shown to be involved in the apoptosis of astrocytes (Iorio et al., 2002), cardiac myocytes (Shneyvays et al., 1998), and HL-60 leukemia cells (Kohno et al., 1996a). As for the truncated P2X7 receptor, reported here for the first time, it remains to be determined if it codes for a functional receptor protein.
While P2X7 receptor mediates the apoptotic action of extracellular ATP in HEK-P2X7 cells and A3 receptor mediates the apoptotic effect in Li-7A cells, their effects on the intracellular mediators of apoptosis, that is, caspase and PARP, are similar. In both cases, there is substantial increase in caspase-3 activation, as evidenced by proteolysis of procaspase-3 and the caspase substrate PARP. Similar findings have been reported for adenosine-induced cell death of human HL-60 leukemia cells, human U-937 lymphoma cells, and mouse neuroblastoma cells (Schrier et al., 2001; 2002). In Li-7A cells, the crucial roles of caspase-3 activation and PARP cleavage in adenosine-induced cell death were supported by evidence that the caspase-3 inhibitor Cl-DEVD.fmk completely abolished adenosine-induced Li-7A cell death.
In summary, we demonstrated that the cytotoxic effect of extracellular ATP and other adenine nucleotides on the human Li-7A hepatoma cells is mostly due to the generation of adenosine by ectonucleotidases. The cytotoxic effect of adenosine is mediated by the A3 adenosine receptor. Ligand binding to the receptor activates caspase-3, leading to apoptosis.
L.T.W. acknowledges the Biology Department, San Diego State University, for supporting the Microchemical Core Facility. This work was also supported by the California Metabolic Research Foundation to A.F.K.