A His-155 to Tyr Polymorphism Confers Gain-of-Function to the Human P2X7 Receptor of Human Leukemic Lymphocytes1
http://www.100md.com
免疫学杂志 2005年第13期
Abstract
The P2X7R is an ATP-gated cation channel expressed in hemopoietic cells that participates in both cell proliferation and apoptosis. Expression and function of the P2X7R have been associated with the clinical course of patients affected by chronic lymphocytic leukemia (CLL). Functional variants causing loss-of-function of the P2X7R have been identified, namely, polymorphisms 1513A>C (E496A), 1729T>A (I568N), and 946G>A (R307Q). Here we investigated other nonsynonymous polymorphisms located either in the extracellular portion of the receptor, such as the 489C>T (H155Y) variant, or in the long cytoplasmic tail of the receptor, such as the 1068G>A (A348T), 1096C>G (T357S), and 1405A>G (Q460R) variants. P2X7R function was monitored by measuring ATP-induced Ca2+ influx in PBL of patients affected by CLL and in recombinant human embryonic kidney (HEK) 293 cells stably transfected with each single P2X7 allelic variant. Ca2+ influx was markedly reduced in association with the 1513C allele, whereas variants located in the same intracellular domain, such as the 1068A, 1096G, or 1405G variants, were associated with a minor functional decrease. Significant Ca2+ flux increase was observed in lymphocytes from CLL patients bearing the 489C/T and 489T/T genotypes in association with the 1513A/A genotype. Functional analysis in recombinant HEK293 cells expressing P2X7R confirmed an increased ATP-dependent activation of the P2X7 489T mutant with respect to the wild type receptor, as assessed by both by [Ca2+]i influx and ethidium uptake experiments. These data identify the 489C>T as a gain-of-function polymorphism of the P2X7R.
Introduction
The P2X7R belongs to a family of ligand-gated channels activated by extracellular ATP (1, 2). The P2X7 subunit is a protein 595 aa long, comprising two transmembrane domains, a bulky extracellular residue and intracellular N and C termini (2, 3, 4). Binding of extracellular ATP activates multimerization of P2XR subunits into trimers and hexamers, allowing transmembrane passage of cations and small molecules (2, 5). Short stimulation with extracellular ATP, or the more potent agonist 2',3'-O-(benzoyl-4-benzoyl)-ATP (Bz-ATP),3 activates Na+, K+, and Ca2+ fluxes through P2X7 (6) and promotes several responses among which is cell proliferation (7). Prolonged or repeated exposure to the ligand induces the formation of a cytolytic pore permeable to large hydrophilic molecules such as ethidium bromide, Lucifer yellow, or Yo-Pro-1 (Molecular Probes) and triggers cell death (1, 8, 9).
An A to C polymorphism at position 1513 (1513A>C) of the coding region of the P2X7 gene, leading to the substitution of Glu-496 to Ala residue in the intracellular C-terminal tail, has been described (10). Monocytes and lymphocytes from a subject carrying the homozygous substitution 1513A>C showed nonfunctional P2X7R, whereas heterozygous status halved the functional response (10). Glu-496 to Ala substitution was not associated with reduction of cell surface expression of P2X7 protein in lymphocytes, and loss-of-function was confirmed in recombinant human embryonic kidney (HEK) 293 cells expressing the mutated Glu-496 to Ala receptor (10). More recently, the Glu-496 to Ala polymorphism was shown to impair ATP-mediated immune responses such as the killing of mycobacteria by human macrophages and the release of IL-1 and IL-18 from human monocytes (11, 12, 13). A T to A polymorphism in position 1729 (1729 T>A) inducing the Ile-568 to Asn change has been shown to reduce P2X7R function by preventing correct intracellular trafficking and localization to the plasma membrane (14). A T to C change in position 1352 (Pro-451 to Leu) has been found to impair cell death in murine thymocytes (15). Moreover, polymorphisms in the putative promoter region have been recently reported (16), and a significant protective association against tuberculosis was found for one of them (17). Many other P2X7 gene variants have been also included in The Single Nucleotide Polymorphism (TSC) Consortium database, but until now the functional effects of these sequence variants have not been completely characterized.
B cell chronic lymphocytic leukemia (CLL) is the most frequent type of adult leukemia in the western world, resulting in the accumulation of CD5+ B lymphocytes in the blood and bone marrow. An indolent variant of the disease is recognized, in that some of the CLL patients have stable blood counts and survive for decades without therapy. In contrast, patients with the aggressive form present a rapid progression of the disease requiring cytotoxic treatment. The P2X7R, which is expressed in hemopoietic cells (18), has been also implicated in the clinical course of patients affected by CLL on the basis of different evidence. Increased expression of the P2X7R in peripheral blood cells of patients affected by CLL has been demonstrated (19). Furthermore, peripheral lymphocytes obtained from CLL patients with the aggressive disease had greater P2X7-dependent Ca2+ influx than those from patients with the indolent variant, suggesting that P2X7 function may directly correlate with the severity of the disease course (19). In a different clinical study performed in Australian individuals, the P2X7 loss-of-function 1513C allele was found to be 3-fold more frequent in patients affected by CLL than in controls (20), thus implying that this polymorphism could be involved in the pathogenesis of this disease. Studies from different groups could not confirm this finding in their respective CLL cohorts (21, 22), and a follow-up study from the Australian group concluded that the increased frequency of the 1513C allele was mainly confined to familial CLL cases (23). Although some discrepancies emerge from these studies (24), the correlation observed between reduced expression and function of P2X7 in ex vivo peripheral blood samples of patients with the indolent CLL variant (19) is suggestive of a role of P2X7 in CLL.
We extended the genetic analysis from the 1513A>C polymorphism to other uncharacterized nonsynonymous P2X7 gene polymorphisms in CLL patients attended at the Section of Hematology, University of Ferrara, Ferrara, Italy. PBL from individual patients were screened for the 1513A>C polymorphism and P2X7-mediated functional responses. P2X7 function was also studied in HEK293 cells expressing the site-directed mutated versions of the P2X7 gene corresponding to the minor alleles of the nonsynonymous polymorphisms. The main observation indicates that the P2X7 489C>T allelic variant, leading to the His 155 into Tyr change in the extracellular domain of the receptor, confers gain of function to the P2X7R.
Materials and Methods
Materials
G418 sulfate was purchased from Calbiochem; HEK293 cells were from American Type Culture Collection; fura 2-AM was from Molecular Probes; Nucleon was from Amersham; TaqMan minor groove binder (MGB) probe and PCR primer oligonucleotides, PCR Master Mix, AmpliTaq Gold DNA polymerase, dNTPs with dUTP, and AmpErase uracil-N-glycosylase (UNG) were from Applied Biosystems; pcDNA3 expression vector was from Invitrogen Life Technologies; DpnI and NotI restriction endonucleases, XL1-Blue cells, and PfuTurbo DNA polymerase were from Stratagene; cell culture medium (RPMI, DMEM, Ham’s F-12), ionomycin, ethidium bromide, sulfinpyrazone, salts and all the other chemicals were from Sigma-Aldrich.
Normal and diseased subjects
A cohort of 62 patients affected by CLL, receiving medical examination at the Section of Hematology of the University of Ferrara from March to December 2003, were included in the P2X7R genotypic and functional analysis. All patients were CD5+CD23+ and were diagnosed as previously reported (19). They had a follow-up of at least 3 years and did not receive cytotoxic therapy for at least 2 mo before blood sampling. Peripheral blood samples from a cohort of 100 healthy subjects were collected in parallel within the Bone Marrow Bank donors at the Hospital of Ferrara (Ferrara, Italy). Approval was obtained from the Institutional Review Board for these studies. Informed consent was provided according to the Declaration of Helsinki.
Cells
Lymphocytes were isolated from peripheral blood samples by Ficoll-Hypaque gradient, washed in RPMI medium and suspended in the appropriate solutions for cytosolic Ca2+ analysis or DNA extraction. HEK293 cells were cultured in DMEM and F-12 medium (1:1) containing 15% heat-inactivated FCS, 100 U/ml penicillin, and 100 μg/ml streptomycin.
Cytoplasmic free Ca2+ concentration measurements
Intracellular Ca2+ concentration ([Ca2+]i) transients were measured with the fluorescent indicator fura 2-AM, as described previously (25). Briefly, in each time course, 1 x 106 cells were loaded for 15 min with 2 μM fura 2-AM and then suspended in a low sodium solution (300 mM sucrose, 1 mM MgCl2, 1 mM K2HPO4, 5 mM KHCO3, 5.5 mM glucose, 1 mM CaCl2, 20 mM HEPES, pH 7.4) to minimize Ca2+/Na+ competition for permeation through P2X7, in the presence of 250 μM sulfinpyrazone, to reduce excretion of intracellular fura-2 (26). Cells were kept at 37°C in a thermostat-controlled and magnetically stirred cuvet of a PerkinElmer LS-50 spectrofluorometer, and [Ca2+]i was determined with the 340/380 nm excitation ratio and at 505 nm emission wavelengths. ATP (1 mM) or Bz-ATP (100 μM) were added to study the activation of P2X7R. The calcium ionophore ionomycin (1 μM) was added at the end of each time course as an internal assay control. The ATP-stimulated [Ca2+]i rise above resting concentration was used as a readout for P2X7R functional response.
Ethidium uptake
Changes in plasma membrane permeability on exposure to ATP were studied by ethidium uptake as previously described (25). Briefly, cells (1 x 106/ml) were kept at 37°C in a thermostat-controlled and magnetically stirred cuvet of a PerkinElmer LS-50 spectrofluorometer in Ca2+-free standard saline solution containing EDTA (125 mM NaCl, 5 mM KCl, 1 mM MgSO4, 2 mM EDTA, 5.5 mM glucose, 1 mM NaH2PO4, 5 mM NaHCO3, 20 mM HEPES, pH 7.4) and in the presence of 20 μM ethidium bromide. Fluorescence changes were acquired at 360 nm and 580 nm excitation and emission wavelengths, respectively. After acquisition of the baseline signal (100 s), ATP (3 mM) was added to monitor P2X7R-dependent ethidium bromide uptake for further 500 s. At the end of each time course, 200 μM digitonin was added to measure the fluorescence increment corresponding to the 100% membrane permeabilization. The percent response to ATP was calculated in each time course by measuring the increment of fluorescence 500 s after addition of ATP in comparison with that obtained upon addition of digitonin.
P2X7 polymorphism analysis
Genomic DNA was extracted and purified by sodium perchlorate and chloroform with a Nucleon commercial kit according to the manufacturer’s instructions. The single nucleotide polymorphisms (SNPs) 489C>T (H155Y), 1068G>A (A348T), 1096C>G (T357S), 1405A>G (Q460R), and 1513A>C (E496A) were analyzed in genomic DNA samples by the TaqMan MGB probe technique. Briefly, 100 ng of genomic DNA were added to the PCR Master mix containing two allele-specific fluorescent probes, a primer pair, AmpliTaq Gold DNA polymerase, dNTPs with dUTP, and AmpErase UNG in a final volume of 20 μl. HPLC-purified oligonucleotides were manufactured by and purchased from Applied Biosystems. TaqMan MGB probes were obtained by covalently linking an allele-specific fluorescent molecule to the 5' end (either VIC or 6-FAM) and a nonfluorescent quencher together with a MGB molecule to the 3' end of each oligonucleotide. Allele-specific hybridization was detected in real time with an ABI PRISM 7000 Sequence Detection System instrument (Applied Biosystems) during 30 amplification cycles at 92 and 60°C. TaqMan MGB probe assays were designed to allow the presence of an internal control of the processes of both amplification and hybridization in each assay tube. Design has been constrained to run the analysis of each SNP under the same two-step thermal cycling conditions. TaqMan MGB probes and primer oligonucleotide sequences are reported in Table I. The 489C>T polymorphism was analyzed using probes and primers of an Assay-on-Demand kit manufactured by Applied Biosystems. Genotype was finally assigned by allelic discrimination with analysis of the fluorescent signal both during and at the end of the template amplification.
Site-directed mutagenesis of human P2X7
The hP2X7p635 plasmid, a NotI-NotI insert of the human wild-type P2X7 coding sequence (GenBank accession number Y09561) within a pcDNA3 expression vector, donated by Dr. Gary Buell (Serono Pharmaceutical Research Institute, Geneva, Switzerland), was amplified into XL1-Blue-competent cells and used for site-directed mutagenesis. Two oligonucleotide primers containing the desired point mutation, each complementary to the opposite strands of the plasmid and symmetrically flanking the base to be mutated, were designed for each polymorphic site as reported in Table II. Mutagenesis primers were extended by temperature cycling to generate a mutated plasmid. To increase mutation efficiency and decrease potential for generating random mutations, the lowest cycle number allowing primer extension and the high fidelity PfuTurbo DNA polymerase were used. Wild-type templates were specifically digested with the DpnI endonuclease, which recognizes only methylated and hemimethylated DNA. Generation of the mutated product and the efficiencies of DpnI digestion and of cell transformation were checked throughout the procedure. All clones were checked by cycle sequencing. Sequencing primers (four direct and four reverse) allowing the analysis of the whole insert from T7 to Sp6 sequences of the pcDNA3 plasmid were designed, to check both the presence of the desired mutation and to exclude unwanted additional sequence variations. Cycle sequencing were conducted on service at the Bio Molecular Research sequencing core of the Centro Ricerche Interdipartimentale Biotecnologie Innovative-University of Padua (Padua, Italy).
Transfection of HEK293 cells
HEK293 cells (2.5 x 106) were transfected with plasmids (20–40 μg) encoding the wild-type or the mutated versions of the human P2X7 sequence by calcium phosphate precipitation, cloned, and kept under selection in the presence of 0.2 mg/ml G418 sulfate (Geneticin) as described (27).
Immunofluorescence and flow cytometry
Surface expression of the P2X7R was performed by indirect immunofluorescence and flow cytometry with a mouse monoclonal anti-P2X7 Ab directed against the extracellular domain of the human P2X7R, donated by Dr. Gary Buell. Primary Ab was incubated for 1 h at 4°C. Cells were washed in saline solution and then incubated with FITC-conjugated anti-mouse Ab for 1 h at 4°C. Fluorescent signal was collected and analyzed with a single argon laser cytofluorometer (BD Biosciences).
Results
Frequencies of P2X7 polymorphisms in CLL patients and healthy controls
We analyzed the allelic frequencies of different nonsynonymous SNPs of the P2X7 gene. As shown in Table III, 36 of the 100 healthy control subjects were 1513A/C heterozygotes, and 2 of them were homozygous for the 1513C allele. The 1513A>C minor allele frequency is in Hardy-Weinberg equilibrium and within the range which was found in the other European reports already published (21, 28, 29, 30). The frequency of the 1513C allele in 62 CLL patients attended at the Section of Hematology of the University of Ferrara was 0.234, and the genotype frequencies were not significantly different from those of the healthy controls (p = 0.574) (Table III). The list of P2X7 SNPs in the TSC database is growing constantly, whereas the allelic frequencies of these SNPs have been rarely determined. We chose to investigate four P2X7 gene nonsynonymous polymorphisms that have been recently reported in the TSC database, namely 489C>T (His-155 to Tyr), 1068G>A (Ala-348 to Thr), 1096C>G (Thr-357 to Ser), 1405A>G (Gln-460 to Arg). A frequency >1% was found in our healthy control sample, the minor allele frequency ranging from a minimum of 0.040 (1096G allele) to a maximum of 0.470 (489T allele), as shown in Table III. The genotype frequencies of each SNP were in Hardy-Weinberg equilibrium, and no significant differences were found between CLL patients and healthy controls (Table III).
P2X7 1513C allele results in loss of Ca2+ influx in lymphocytes ex vivo
The loss of P2X7R function caused by the 1513C allele was originally demonstrated by measuring ATP-induced ethidium uptake in lymphocytes of CLL patients (10). We studied P2X7R function in PBL purified from 62 CLL patients by measuring the rise of [Ca2+]i dependent on ATP-induced Ca2+ influx through P2X7R, as previously described (19). The difference () between peak and resting [Ca2+]i has been plotted in relation to the 1513A>C genotype of the CLL patients, as shown in Fig. 1A. The average [Ca2+]i on exposure to ATP in CLL patients was 133 ± 11 nM (mean ± SEM, n = 62). CLL patients who were homozygous at the 1513 (C/C) polymorphism had a marked reduction of the average P2X7 function ([Ca2+]i) in comparison with those bearing the 1513A/A genotype (31 ± 10 vs 170 ± 15, mean ± SEM, p < 0.01 by Student’s t test for unpaired data). The 1513A/C-heterozygous patients also presented a significant functional reduction as compared with 1513A/A patients (p < 0.01 by Student’s t test for unpaired data), the average [Ca2+]i being at an intermediate level between that of 1513A/A and 1513C/C patients, as shown in Fig. 1A. Thus, in agreement with results obtained with ATP-dependent ethidium uptake by other investigators (10, 20), the data confirm that the P2X7 1513C allele confers a loss of P2X7R function.
Correlation between P2X7R function and polymorphisms ex vivo
The P2X7 489C>T, 1068G>A, 1096C>G, 1405A>G polymorphisms, which code for changes in P2X7R primary amino acid sequence, could either be neutral or modulate P2X7R function. Thus we correlated the P2X7R function measured in lymphocytes from CLL patients ex vivo with the genotype of each of these polymorphisms. Because the specific functional modulation of each of these polymorphisms could be disguised by the presence of the 1513C loss-of-function allele, we analyzed the role of each SNP in the subset of CLL patients bearing the 1513A/A genotype. Fig. 1B shows P2X7R activation in association with the 489T allele, coding for the His-155 to Tyr change in the extracellular domain of the receptor involved in ATP binding. A significant gain of function was observed in lymphocytes of CLL patients bearing the 489T allele, because the average [Ca2+]i is greater in the presence of the 489C/T and the 489T/T genotypes with respect to the 489C/C one (C/C vs C/T, p < 0.05; C/C vs T/T, p < 0.05, by Student’s t test for unpaired data). The association between P2X7R activation and the other polymorphisms located in the intracellular C-terminal tail close to the loss-of-function allele 1513C is reported in Fig. 2. The 1068G/A and 1068G/G genotypes are associated with a slight decrease of P2X7R function with respect to the 1068G/G genotype, albeit this is not statistically significant (p = 0.300, by Student’s t test for unpaired data), as reported in Fig. 2A. Minimal variation of the P2X7R activation was observed in lymphocytes from CLL patients bearing the 1096C/G and 1405A/G genotypes in comparison with those bearing the most frequent 1096C/C and 1405A/A genotypes (p = 0.546 and p = 0.673, respectively, by Student’s t test for unpaired data), as shown in Fig. 2, B and C. When correlation between P2X7 function and genotype was investigated in the whole sample of 62 CLL patients, therefore including also the subjects bearing the 1513C loss-of-function allele, no significant differences were observed (data not shown). In particular, the average gain of function observed in lymphocytes from the subset of CLL patients bearing the 489C/T and 489T/T genotypes in association with the 1513A/A genotype was abolished.
Expression and function of P2X7 mutants in vitro
The effect of the P2X7 genetic variants on receptor function was next studied in vitro by expressing of each mutated version of the human P2X7 gene in the HEK293 cell line. The rise of [Ca2+]i upon Bz-ATP-induced Ca2+ influx through the activated P2X7R was measured, and [Ca2+]i transients of each single mutant were compared with those obtained upon transfection of the P2X7 gene version coding the major alleles of each SNP (P2X7 wild type). As shown in Fig. 3A, the high affinity P2X7R ligand Bz-ATP induces a rapid rise of [Ca2+]i in recombinant HEK293 cells expressing the wild-type P2X7 gene (HEK293 P2X7 wild type), whereas no response is observed in the cell line transfected with the insertless expression vector pcDNA3 (HEK293 pcDNA3), confirming the absence of constitutive expression of the P2X7R in the parental HEK293 cell line. The calcium ionophore ionomycin was added at the end of each time course as a positive control of the assay system. No activation of the P2X7R was observed in the majority of HEK293 cells expressing the P2X7 1513C mutant (HEK293 P2X7 1513C), as shown in Fig. 3B, although we observed a slight [Ca2+]i rise in few cell clones hyperexpressing the P2X7 1513C mutant, in agreement with the original observation obtained by Gu et al. (10) by ethidium uptake in HEK293 cells in vitro and lymphomonocytes ex vivo. Bz-ATP-dependent activation of the P2X7 1068A, 1096G, and 1405G mutants was very similar to the response of the P2X7 wild type receptor, the only mutant showing a slightly reduced response being 1068A, as shown in the time courses of Fig. 3C. The P2X7 489T mutant receptor showed an higher functional activation compared with P2X7 wild type, as reported in Fig. 4A. Surface expression levels of the wild-type and 489T mutant receptors were similar, as shown by the FACS reported in Fig. 4B, excluding that the greater functional response of the 489T mutant is due to higher protein surface expression levels. Under these conditions of overlapping flow cytometry profiles, average [Ca2+]i of HEK293 cells expressing the P2X7 wild-type or the 489T mutant receptor was 132 ± 20 and 196 ± 22 nM (mean ± SEM, from five separate experiments; p < 0.01 by Student’s t test for paired data), respectively. Analysis of P2X7 489T mutant receptor permeability was extended to larger cations such as ethidium. Recombinant HEK293 cells were incubated in the presence of extracellular ethidium bromide and, after reaching a stable baseline, exposed to ATP, and the rate of ethidium uptake was measured. Response to extracellular ATP was calculated as percent of permeability increase caused by digitonin, as described in Materials and Methods. In clones with similar surface expression levels of the P2X7R (Fig. 5B), the rate of ethidium uptake was higher in the P2X7 489T than in the wild-type HEK293 cells, as shown in the representative time course of Fig. 5A. In these experiments, the P2X7 1513C HEK293 cell line showed a very low ATP-dependent ethidium uptake (Fig. 5A). The average percent permeability increase of the P2X7 wild-type HEK293 cells was 30.1 ± 0.3, whereas that of P2X7 489T HEK293 cells was 44.4 ± 4.4 (mean ± SEM, from four separate experiments; p < 0.05 by Student’s t test for paired data). Therefore, the functional analyses of P2X7R mutants in vitro indicate that the P2X7 489T allele confers a significant gain of receptor function, that the 1068A variant causes a slight functional reduction, and that the 1096G and 1405G variants have an almost neutral effect.
Discussion
The main finding of this functional study in ex vivo lymphocytes from CLL patients and in in vitro recombinant mutant cell lines is that the 489C>T variant of the P2X7 gene is associated with gain-of-function of the receptor, that the 1068G>A variant slightly reduces receptor activation, and that the 1096C>G and 1405A>G variants have an almost neutral effect. In addition, we confirm that the 1513A>C polymorphism causes loss-of-function of the receptor, and we report that the prevalence of these five variants of the P2X7 gene is similar in patients affected by CLL and healthy control individuals.
A dual role of the P2X7R is widely accepted (24), because it promotes cell apoptosis upon prolonged exposure to high ATP concentrations, whereas at tonic, low concentrations of the ligand, it induces cell growth (7). The modulation of P2X7R function is attracting interest for its potential implications in human pathology, and in particular, inheritance of a defective P2X7R has been claimed to be associated to predisposition to CLL. In a study conducted in an Australian sample population, the prevalence of the P2X7 1513A>C loss-of-function polymorphism was 3-fold greater in patients with CLL than in controls (20). Wiley et al. suggested that inheritance of the 1513C allele could contribute to the pathogenesis of CLL, as in the presence of this allele, P2X7-dependent apoptosis is impaired, and this could result in accumulation of neoplastic B cells. Accordingly, they proposed P2X7 as a tumor suppressor gene. However, subsequent studies performed in CLL patients from Sweden, the United Kingdom, Germany (28); extension of the original paper in familial cases (21, 23, 29, 30); and the present study create a consensus on the observation that the P2X7 1513C allele prevalence is not increased in the sporadic form of CLL, which makes less likely the direct involvement of the P2X7 1513A>C polymorphism in the pathogenesis of CLL.
Truncation or partial deletions of the P2X7 protein in the C-terminal protein tail impair the ATP-dependent transmembrane ion transport, indicating that different domains within the C-terminal end of the receptor either facilitate the assembly of the monomers into the channel complex or are relevant to target the receptor in the plasma membrane (1, 2, 9). Although extended deletions substantially modify P2X7R activation, single residue changes within the same domain have unpredictable effects. The 1513A>C sequence variant, encoding the Glu-496 to Ala change, was found to impair both Ba2+ and ethidium uptake in blood peripheral mononuclear and transiently transfected recombinant HEK293 cells (10). By studying P2X7R-dependent intracellular Ca2+ rise, we confirm that the 1513C allele results in loss-of-function of the receptor both in lymphocytes of CLL patients ex vivo and in recombinant HEK293 cells in vitro (Figs. 1 and 3). Our data do not provide insights on the mechanism responsible for this functional defect, albeit it might be speculated that the Glu to Ala residue change might weaken the electrostatic interactions during the assembly of the P2X7 channel complex in the plasma membrane (10). Here we investigated three additional allelic variants changing the sequence of the C-terminal protein tail, namely the amino acid changes Ala-348 to Thr (1068G>A), Thr-357 to Ser (1096C>G), and Gln-460 to Arg (1405A>G), which are in proximity to the Glu-496 to Ala variation. The functional analysis of these amino acid changes conducted both in lymphocytes from CLL patients and in P2X7-transfected HEK293 cells indicate a slight reduction (1068A) or neutral effect (1096G and 1405G) on receptor activation (Figs. 2 and 3), although the possibility remains open that combined residue changes in the P2X7R could alter channel activation more profoundly.
Little is known about the pocket allowing ATP binding to the extracellular loop of the receptor. Positively charged residues, which are candidate binding sites for ATP, have been mutated and expressed in recombinant cell models, and both Lys-193 and Lys-311 were shown to be essential in ATP binding to the extracellular loop of the human P2X7R (31). More recently, the single nucleotide substitution 946G>A, encoding the Arg-307 to Gln change, has been identified in Caucasian subjects with an allelic frequency of 0.014 (32). Interestingly, P2X7R-dependent Ba2+, Rb+, and ethidium+ influx was markedly reduced in PBL and monocytes obtained from 946G/A heterozygous individuals, whereas P2X7R cell surface expression was conserved, indicating that Arg-307 is critical for ATP binding and receptor activation (32). In this study, we found that the frequent 489C>T allelic variant encoding the change of His-155 to Tyr is associated with an increased P2X7R-dependent Ca2+ influx in transfected HEK293 cells (Fig. 4). Interestingly, gain of function associated with the 489T allele was also observed in PBL from 1513A/A homozygous CLL patients (Fig. 1) but not in 1513A/C and 1513C/C subjects, indicating that the loss- of-function conferred by the E496A residue change in the long C-terminal tail cannot be recovered by the H155Y substitution in the extracellular loop of the receptor. As the list of P2X7 allelic variants conferring loss of receptor function is not thoroughly defined, it could be possible that the combined presence of other unknown loss-of-function alterations on the same copy of the P2X7 gene might reduce or abolish the gain of function conferred by the 489T allele in blood PBL ex vivo. Thus, the finding of the 489T allele in CLL individuals is expected to increase P2X7R function in peripheral blood lymphocytes to variable extents, depending on the combined presence of loss-of-function variants. How might the change of His-155 to Tyr in the extracellular loop of the receptor increase the P2X7R channel function? Considering the high frequency of the 489T minor allele that we observed in CLL patients and healthy control individuals, a large fraction of the human P2X7 protein should have a Tyr residue in position 155 of the primary sequence. Extensive molecular studies have been performed on the role of 19 Tyr residues of rat P2X7R, both on activation and on the association of P2X7R with accessory proteins, as a function of the phosphorylated status (33), whereas studies on the residue homologous to the H115Y variation of the human sequence protein are missing. The most straightforward explanation might take into consideration that the H155Y change increases the affinity of the agonist to the receptor, although the ATP and Bz-ATP used in our experimental conditions were at saturating concentrations. Considering the change of electrostatic charges consequent to an His to Tyr substitution and the potential of Tyr phosphorylation, alternative speculations could include a facilitation in the recruitment of the P2X7 monomers in the process of assembly of the receptor complexes or a stabilization of the active complexes themselves, thus increasing the ratio between active complexes and inactive monomers. Additional experiments are needed to fully clarify this issue.
Acknowledgments
We thank Gary Buell (Serono Pharmaceutical Research Institute, Geneva, Switzerland), for providing the hP2X7p635 plasmid and the anti-P2X7R mAb, and Richard Rosenquist (Rudbeck Laboratory, University of Uppsala, Uppsala, Sweden) for thoughtful discussions.
Disclosures
The authors have no financial conflict of interest.
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 Ministry of Education and Scientific Research (grants Cofin and Firb), Italian Association for Cancer Research (AIRC) and institutional funds from the University of Ferrara.
2 Address correspondence and reprint requests to Dr. Giulio Cabrini, Laboratorio Patologia Molecolare, Centro Fibrosi Cistica, Azienda Ospedaliera Verona, Piazzale Stefani 1, 37126 Verona, Italy. E-mail address: giulio.cabrini@azosp.vr.it
3 Abbreviations used in this paper: Bz-ATP, 2',3'-O-(benzoyl-4-benzoyl)-ATP; HEK, human embryonic kidney; TSC, The Single Nucleotide Polymorphism Consortium; CLL, chronic lymphocytic leukemia; [Ca2+]i, intracellular Ca2+ concentration; UNG, uracil-N-glycosylase; MGB, minor groove binder; SNP, single nucleotide polymorphism.
Received for publication December 23, 2004. Accepted for publication April 12, 2005.
References
Surprenant, A., F. Rassendren, E. Kawashima, R. A. North, G. Buell. 1996. The cytolytic P2Z receptor for extracellular ATP identified as a P2X receptor (P2X7). Science 272: 735-738.
North, R. A.. 2002. Molecular physiology of P2X receptors. Physiol. Rev. 82: 1013-1067.
Rassendren, F., G. N. Buell, C. Virginio, G. Collo, R. A. North, A. Surprenant. 1997. The permeabilizing ATP receptor, P2X7, cloning and expression of a human cDNA. J. Biol. Chem. 272: 5482-5486
Buell, G., F. Talabot, A. Gos, J. Lorenz, E. Lai, M. A. Morris, S. E. Antonorakis. 1998. Gene structure and chromosomal localization of the human P2X7 receptor. Receptor Channel 5: 347-354.
Nicke, A., H. G. B?umert, J. Rettinger, A. Eichele, G. Lambrecht, E. Mutschler, G. Schmalzing. 1998. P2X1 and P2X3 receptors form stable trimers: a novel structural motif of ligand-gated ion channels. EMBO J. 17: 3016-3028.
Virginio, C., A. MacKenzie, R. A. North, A. Surprenant. 1999. Kinetics of cell lysis, dye uptake and permeability changes in cells expressing the rat P2X7 receptor. J. Physiol. 519: 335-346.
Baricordi, O. R., L. Melchiorri, E. Adinolfi, S. Falzoni, P. Chiozzi, G. Buell, F. Di Virgilio. 1999. Increased proliferation rate of lymphoid cells transfected with the P2X7 ATP receptor. J. Biol. Chem. 274: 33206-33208.
Nuttle, L. C., G. R. Dubyak. 1994. Differential activation of cation channels and nonselective pores by macrophage P2Z purinergic receptors expressed in Xenopus oocytes. J. Biol. Chem. 269: 13988-13999.
Smart, M. L., B. J. Gu, R.G. Panchal, J. S. Wiley, B. Cromer, D. A. Williams, S. Petrou. 2003. P2X7 receptor cell surface expression and cytolytic pore formation are regulated by a distal C-terminal region. J. Biol. Chem. 278: 8853-8860.(Giulio Cabrini2,*, Simone)
The P2X7R is an ATP-gated cation channel expressed in hemopoietic cells that participates in both cell proliferation and apoptosis. Expression and function of the P2X7R have been associated with the clinical course of patients affected by chronic lymphocytic leukemia (CLL). Functional variants causing loss-of-function of the P2X7R have been identified, namely, polymorphisms 1513A>C (E496A), 1729T>A (I568N), and 946G>A (R307Q). Here we investigated other nonsynonymous polymorphisms located either in the extracellular portion of the receptor, such as the 489C>T (H155Y) variant, or in the long cytoplasmic tail of the receptor, such as the 1068G>A (A348T), 1096C>G (T357S), and 1405A>G (Q460R) variants. P2X7R function was monitored by measuring ATP-induced Ca2+ influx in PBL of patients affected by CLL and in recombinant human embryonic kidney (HEK) 293 cells stably transfected with each single P2X7 allelic variant. Ca2+ influx was markedly reduced in association with the 1513C allele, whereas variants located in the same intracellular domain, such as the 1068A, 1096G, or 1405G variants, were associated with a minor functional decrease. Significant Ca2+ flux increase was observed in lymphocytes from CLL patients bearing the 489C/T and 489T/T genotypes in association with the 1513A/A genotype. Functional analysis in recombinant HEK293 cells expressing P2X7R confirmed an increased ATP-dependent activation of the P2X7 489T mutant with respect to the wild type receptor, as assessed by both by [Ca2+]i influx and ethidium uptake experiments. These data identify the 489C>T as a gain-of-function polymorphism of the P2X7R.
Introduction
The P2X7R belongs to a family of ligand-gated channels activated by extracellular ATP (1, 2). The P2X7 subunit is a protein 595 aa long, comprising two transmembrane domains, a bulky extracellular residue and intracellular N and C termini (2, 3, 4). Binding of extracellular ATP activates multimerization of P2XR subunits into trimers and hexamers, allowing transmembrane passage of cations and small molecules (2, 5). Short stimulation with extracellular ATP, or the more potent agonist 2',3'-O-(benzoyl-4-benzoyl)-ATP (Bz-ATP),3 activates Na+, K+, and Ca2+ fluxes through P2X7 (6) and promotes several responses among which is cell proliferation (7). Prolonged or repeated exposure to the ligand induces the formation of a cytolytic pore permeable to large hydrophilic molecules such as ethidium bromide, Lucifer yellow, or Yo-Pro-1 (Molecular Probes) and triggers cell death (1, 8, 9).
An A to C polymorphism at position 1513 (1513A>C) of the coding region of the P2X7 gene, leading to the substitution of Glu-496 to Ala residue in the intracellular C-terminal tail, has been described (10). Monocytes and lymphocytes from a subject carrying the homozygous substitution 1513A>C showed nonfunctional P2X7R, whereas heterozygous status halved the functional response (10). Glu-496 to Ala substitution was not associated with reduction of cell surface expression of P2X7 protein in lymphocytes, and loss-of-function was confirmed in recombinant human embryonic kidney (HEK) 293 cells expressing the mutated Glu-496 to Ala receptor (10). More recently, the Glu-496 to Ala polymorphism was shown to impair ATP-mediated immune responses such as the killing of mycobacteria by human macrophages and the release of IL-1 and IL-18 from human monocytes (11, 12, 13). A T to A polymorphism in position 1729 (1729 T>A) inducing the Ile-568 to Asn change has been shown to reduce P2X7R function by preventing correct intracellular trafficking and localization to the plasma membrane (14). A T to C change in position 1352 (Pro-451 to Leu) has been found to impair cell death in murine thymocytes (15). Moreover, polymorphisms in the putative promoter region have been recently reported (16), and a significant protective association against tuberculosis was found for one of them (17). Many other P2X7 gene variants have been also included in The Single Nucleotide Polymorphism (TSC) Consortium database, but until now the functional effects of these sequence variants have not been completely characterized.
B cell chronic lymphocytic leukemia (CLL) is the most frequent type of adult leukemia in the western world, resulting in the accumulation of CD5+ B lymphocytes in the blood and bone marrow. An indolent variant of the disease is recognized, in that some of the CLL patients have stable blood counts and survive for decades without therapy. In contrast, patients with the aggressive form present a rapid progression of the disease requiring cytotoxic treatment. The P2X7R, which is expressed in hemopoietic cells (18), has been also implicated in the clinical course of patients affected by CLL on the basis of different evidence. Increased expression of the P2X7R in peripheral blood cells of patients affected by CLL has been demonstrated (19). Furthermore, peripheral lymphocytes obtained from CLL patients with the aggressive disease had greater P2X7-dependent Ca2+ influx than those from patients with the indolent variant, suggesting that P2X7 function may directly correlate with the severity of the disease course (19). In a different clinical study performed in Australian individuals, the P2X7 loss-of-function 1513C allele was found to be 3-fold more frequent in patients affected by CLL than in controls (20), thus implying that this polymorphism could be involved in the pathogenesis of this disease. Studies from different groups could not confirm this finding in their respective CLL cohorts (21, 22), and a follow-up study from the Australian group concluded that the increased frequency of the 1513C allele was mainly confined to familial CLL cases (23). Although some discrepancies emerge from these studies (24), the correlation observed between reduced expression and function of P2X7 in ex vivo peripheral blood samples of patients with the indolent CLL variant (19) is suggestive of a role of P2X7 in CLL.
We extended the genetic analysis from the 1513A>C polymorphism to other uncharacterized nonsynonymous P2X7 gene polymorphisms in CLL patients attended at the Section of Hematology, University of Ferrara, Ferrara, Italy. PBL from individual patients were screened for the 1513A>C polymorphism and P2X7-mediated functional responses. P2X7 function was also studied in HEK293 cells expressing the site-directed mutated versions of the P2X7 gene corresponding to the minor alleles of the nonsynonymous polymorphisms. The main observation indicates that the P2X7 489C>T allelic variant, leading to the His 155 into Tyr change in the extracellular domain of the receptor, confers gain of function to the P2X7R.
Materials and Methods
Materials
G418 sulfate was purchased from Calbiochem; HEK293 cells were from American Type Culture Collection; fura 2-AM was from Molecular Probes; Nucleon was from Amersham; TaqMan minor groove binder (MGB) probe and PCR primer oligonucleotides, PCR Master Mix, AmpliTaq Gold DNA polymerase, dNTPs with dUTP, and AmpErase uracil-N-glycosylase (UNG) were from Applied Biosystems; pcDNA3 expression vector was from Invitrogen Life Technologies; DpnI and NotI restriction endonucleases, XL1-Blue cells, and PfuTurbo DNA polymerase were from Stratagene; cell culture medium (RPMI, DMEM, Ham’s F-12), ionomycin, ethidium bromide, sulfinpyrazone, salts and all the other chemicals were from Sigma-Aldrich.
Normal and diseased subjects
A cohort of 62 patients affected by CLL, receiving medical examination at the Section of Hematology of the University of Ferrara from March to December 2003, were included in the P2X7R genotypic and functional analysis. All patients were CD5+CD23+ and were diagnosed as previously reported (19). They had a follow-up of at least 3 years and did not receive cytotoxic therapy for at least 2 mo before blood sampling. Peripheral blood samples from a cohort of 100 healthy subjects were collected in parallel within the Bone Marrow Bank donors at the Hospital of Ferrara (Ferrara, Italy). Approval was obtained from the Institutional Review Board for these studies. Informed consent was provided according to the Declaration of Helsinki.
Cells
Lymphocytes were isolated from peripheral blood samples by Ficoll-Hypaque gradient, washed in RPMI medium and suspended in the appropriate solutions for cytosolic Ca2+ analysis or DNA extraction. HEK293 cells were cultured in DMEM and F-12 medium (1:1) containing 15% heat-inactivated FCS, 100 U/ml penicillin, and 100 μg/ml streptomycin.
Cytoplasmic free Ca2+ concentration measurements
Intracellular Ca2+ concentration ([Ca2+]i) transients were measured with the fluorescent indicator fura 2-AM, as described previously (25). Briefly, in each time course, 1 x 106 cells were loaded for 15 min with 2 μM fura 2-AM and then suspended in a low sodium solution (300 mM sucrose, 1 mM MgCl2, 1 mM K2HPO4, 5 mM KHCO3, 5.5 mM glucose, 1 mM CaCl2, 20 mM HEPES, pH 7.4) to minimize Ca2+/Na+ competition for permeation through P2X7, in the presence of 250 μM sulfinpyrazone, to reduce excretion of intracellular fura-2 (26). Cells were kept at 37°C in a thermostat-controlled and magnetically stirred cuvet of a PerkinElmer LS-50 spectrofluorometer, and [Ca2+]i was determined with the 340/380 nm excitation ratio and at 505 nm emission wavelengths. ATP (1 mM) or Bz-ATP (100 μM) were added to study the activation of P2X7R. The calcium ionophore ionomycin (1 μM) was added at the end of each time course as an internal assay control. The ATP-stimulated [Ca2+]i rise above resting concentration was used as a readout for P2X7R functional response.
Ethidium uptake
Changes in plasma membrane permeability on exposure to ATP were studied by ethidium uptake as previously described (25). Briefly, cells (1 x 106/ml) were kept at 37°C in a thermostat-controlled and magnetically stirred cuvet of a PerkinElmer LS-50 spectrofluorometer in Ca2+-free standard saline solution containing EDTA (125 mM NaCl, 5 mM KCl, 1 mM MgSO4, 2 mM EDTA, 5.5 mM glucose, 1 mM NaH2PO4, 5 mM NaHCO3, 20 mM HEPES, pH 7.4) and in the presence of 20 μM ethidium bromide. Fluorescence changes were acquired at 360 nm and 580 nm excitation and emission wavelengths, respectively. After acquisition of the baseline signal (100 s), ATP (3 mM) was added to monitor P2X7R-dependent ethidium bromide uptake for further 500 s. At the end of each time course, 200 μM digitonin was added to measure the fluorescence increment corresponding to the 100% membrane permeabilization. The percent response to ATP was calculated in each time course by measuring the increment of fluorescence 500 s after addition of ATP in comparison with that obtained upon addition of digitonin.
P2X7 polymorphism analysis
Genomic DNA was extracted and purified by sodium perchlorate and chloroform with a Nucleon commercial kit according to the manufacturer’s instructions. The single nucleotide polymorphisms (SNPs) 489C>T (H155Y), 1068G>A (A348T), 1096C>G (T357S), 1405A>G (Q460R), and 1513A>C (E496A) were analyzed in genomic DNA samples by the TaqMan MGB probe technique. Briefly, 100 ng of genomic DNA were added to the PCR Master mix containing two allele-specific fluorescent probes, a primer pair, AmpliTaq Gold DNA polymerase, dNTPs with dUTP, and AmpErase UNG in a final volume of 20 μl. HPLC-purified oligonucleotides were manufactured by and purchased from Applied Biosystems. TaqMan MGB probes were obtained by covalently linking an allele-specific fluorescent molecule to the 5' end (either VIC or 6-FAM) and a nonfluorescent quencher together with a MGB molecule to the 3' end of each oligonucleotide. Allele-specific hybridization was detected in real time with an ABI PRISM 7000 Sequence Detection System instrument (Applied Biosystems) during 30 amplification cycles at 92 and 60°C. TaqMan MGB probe assays were designed to allow the presence of an internal control of the processes of both amplification and hybridization in each assay tube. Design has been constrained to run the analysis of each SNP under the same two-step thermal cycling conditions. TaqMan MGB probes and primer oligonucleotide sequences are reported in Table I. The 489C>T polymorphism was analyzed using probes and primers of an Assay-on-Demand kit manufactured by Applied Biosystems. Genotype was finally assigned by allelic discrimination with analysis of the fluorescent signal both during and at the end of the template amplification.
Site-directed mutagenesis of human P2X7
The hP2X7p635 plasmid, a NotI-NotI insert of the human wild-type P2X7 coding sequence (GenBank accession number Y09561) within a pcDNA3 expression vector, donated by Dr. Gary Buell (Serono Pharmaceutical Research Institute, Geneva, Switzerland), was amplified into XL1-Blue-competent cells and used for site-directed mutagenesis. Two oligonucleotide primers containing the desired point mutation, each complementary to the opposite strands of the plasmid and symmetrically flanking the base to be mutated, were designed for each polymorphic site as reported in Table II. Mutagenesis primers were extended by temperature cycling to generate a mutated plasmid. To increase mutation efficiency and decrease potential for generating random mutations, the lowest cycle number allowing primer extension and the high fidelity PfuTurbo DNA polymerase were used. Wild-type templates were specifically digested with the DpnI endonuclease, which recognizes only methylated and hemimethylated DNA. Generation of the mutated product and the efficiencies of DpnI digestion and of cell transformation were checked throughout the procedure. All clones were checked by cycle sequencing. Sequencing primers (four direct and four reverse) allowing the analysis of the whole insert from T7 to Sp6 sequences of the pcDNA3 plasmid were designed, to check both the presence of the desired mutation and to exclude unwanted additional sequence variations. Cycle sequencing were conducted on service at the Bio Molecular Research sequencing core of the Centro Ricerche Interdipartimentale Biotecnologie Innovative-University of Padua (Padua, Italy).
Transfection of HEK293 cells
HEK293 cells (2.5 x 106) were transfected with plasmids (20–40 μg) encoding the wild-type or the mutated versions of the human P2X7 sequence by calcium phosphate precipitation, cloned, and kept under selection in the presence of 0.2 mg/ml G418 sulfate (Geneticin) as described (27).
Immunofluorescence and flow cytometry
Surface expression of the P2X7R was performed by indirect immunofluorescence and flow cytometry with a mouse monoclonal anti-P2X7 Ab directed against the extracellular domain of the human P2X7R, donated by Dr. Gary Buell. Primary Ab was incubated for 1 h at 4°C. Cells were washed in saline solution and then incubated with FITC-conjugated anti-mouse Ab for 1 h at 4°C. Fluorescent signal was collected and analyzed with a single argon laser cytofluorometer (BD Biosciences).
Results
Frequencies of P2X7 polymorphisms in CLL patients and healthy controls
We analyzed the allelic frequencies of different nonsynonymous SNPs of the P2X7 gene. As shown in Table III, 36 of the 100 healthy control subjects were 1513A/C heterozygotes, and 2 of them were homozygous for the 1513C allele. The 1513A>C minor allele frequency is in Hardy-Weinberg equilibrium and within the range which was found in the other European reports already published (21, 28, 29, 30). The frequency of the 1513C allele in 62 CLL patients attended at the Section of Hematology of the University of Ferrara was 0.234, and the genotype frequencies were not significantly different from those of the healthy controls (p = 0.574) (Table III). The list of P2X7 SNPs in the TSC database is growing constantly, whereas the allelic frequencies of these SNPs have been rarely determined. We chose to investigate four P2X7 gene nonsynonymous polymorphisms that have been recently reported in the TSC database, namely 489C>T (His-155 to Tyr), 1068G>A (Ala-348 to Thr), 1096C>G (Thr-357 to Ser), 1405A>G (Gln-460 to Arg). A frequency >1% was found in our healthy control sample, the minor allele frequency ranging from a minimum of 0.040 (1096G allele) to a maximum of 0.470 (489T allele), as shown in Table III. The genotype frequencies of each SNP were in Hardy-Weinberg equilibrium, and no significant differences were found between CLL patients and healthy controls (Table III).
P2X7 1513C allele results in loss of Ca2+ influx in lymphocytes ex vivo
The loss of P2X7R function caused by the 1513C allele was originally demonstrated by measuring ATP-induced ethidium uptake in lymphocytes of CLL patients (10). We studied P2X7R function in PBL purified from 62 CLL patients by measuring the rise of [Ca2+]i dependent on ATP-induced Ca2+ influx through P2X7R, as previously described (19). The difference () between peak and resting [Ca2+]i has been plotted in relation to the 1513A>C genotype of the CLL patients, as shown in Fig. 1A. The average [Ca2+]i on exposure to ATP in CLL patients was 133 ± 11 nM (mean ± SEM, n = 62). CLL patients who were homozygous at the 1513 (C/C) polymorphism had a marked reduction of the average P2X7 function ([Ca2+]i) in comparison with those bearing the 1513A/A genotype (31 ± 10 vs 170 ± 15, mean ± SEM, p < 0.01 by Student’s t test for unpaired data). The 1513A/C-heterozygous patients also presented a significant functional reduction as compared with 1513A/A patients (p < 0.01 by Student’s t test for unpaired data), the average [Ca2+]i being at an intermediate level between that of 1513A/A and 1513C/C patients, as shown in Fig. 1A. Thus, in agreement with results obtained with ATP-dependent ethidium uptake by other investigators (10, 20), the data confirm that the P2X7 1513C allele confers a loss of P2X7R function.
Correlation between P2X7R function and polymorphisms ex vivo
The P2X7 489C>T, 1068G>A, 1096C>G, 1405A>G polymorphisms, which code for changes in P2X7R primary amino acid sequence, could either be neutral or modulate P2X7R function. Thus we correlated the P2X7R function measured in lymphocytes from CLL patients ex vivo with the genotype of each of these polymorphisms. Because the specific functional modulation of each of these polymorphisms could be disguised by the presence of the 1513C loss-of-function allele, we analyzed the role of each SNP in the subset of CLL patients bearing the 1513A/A genotype. Fig. 1B shows P2X7R activation in association with the 489T allele, coding for the His-155 to Tyr change in the extracellular domain of the receptor involved in ATP binding. A significant gain of function was observed in lymphocytes of CLL patients bearing the 489T allele, because the average [Ca2+]i is greater in the presence of the 489C/T and the 489T/T genotypes with respect to the 489C/C one (C/C vs C/T, p < 0.05; C/C vs T/T, p < 0.05, by Student’s t test for unpaired data). The association between P2X7R activation and the other polymorphisms located in the intracellular C-terminal tail close to the loss-of-function allele 1513C is reported in Fig. 2. The 1068G/A and 1068G/G genotypes are associated with a slight decrease of P2X7R function with respect to the 1068G/G genotype, albeit this is not statistically significant (p = 0.300, by Student’s t test for unpaired data), as reported in Fig. 2A. Minimal variation of the P2X7R activation was observed in lymphocytes from CLL patients bearing the 1096C/G and 1405A/G genotypes in comparison with those bearing the most frequent 1096C/C and 1405A/A genotypes (p = 0.546 and p = 0.673, respectively, by Student’s t test for unpaired data), as shown in Fig. 2, B and C. When correlation between P2X7 function and genotype was investigated in the whole sample of 62 CLL patients, therefore including also the subjects bearing the 1513C loss-of-function allele, no significant differences were observed (data not shown). In particular, the average gain of function observed in lymphocytes from the subset of CLL patients bearing the 489C/T and 489T/T genotypes in association with the 1513A/A genotype was abolished.
Expression and function of P2X7 mutants in vitro
The effect of the P2X7 genetic variants on receptor function was next studied in vitro by expressing of each mutated version of the human P2X7 gene in the HEK293 cell line. The rise of [Ca2+]i upon Bz-ATP-induced Ca2+ influx through the activated P2X7R was measured, and [Ca2+]i transients of each single mutant were compared with those obtained upon transfection of the P2X7 gene version coding the major alleles of each SNP (P2X7 wild type). As shown in Fig. 3A, the high affinity P2X7R ligand Bz-ATP induces a rapid rise of [Ca2+]i in recombinant HEK293 cells expressing the wild-type P2X7 gene (HEK293 P2X7 wild type), whereas no response is observed in the cell line transfected with the insertless expression vector pcDNA3 (HEK293 pcDNA3), confirming the absence of constitutive expression of the P2X7R in the parental HEK293 cell line. The calcium ionophore ionomycin was added at the end of each time course as a positive control of the assay system. No activation of the P2X7R was observed in the majority of HEK293 cells expressing the P2X7 1513C mutant (HEK293 P2X7 1513C), as shown in Fig. 3B, although we observed a slight [Ca2+]i rise in few cell clones hyperexpressing the P2X7 1513C mutant, in agreement with the original observation obtained by Gu et al. (10) by ethidium uptake in HEK293 cells in vitro and lymphomonocytes ex vivo. Bz-ATP-dependent activation of the P2X7 1068A, 1096G, and 1405G mutants was very similar to the response of the P2X7 wild type receptor, the only mutant showing a slightly reduced response being 1068A, as shown in the time courses of Fig. 3C. The P2X7 489T mutant receptor showed an higher functional activation compared with P2X7 wild type, as reported in Fig. 4A. Surface expression levels of the wild-type and 489T mutant receptors were similar, as shown by the FACS reported in Fig. 4B, excluding that the greater functional response of the 489T mutant is due to higher protein surface expression levels. Under these conditions of overlapping flow cytometry profiles, average [Ca2+]i of HEK293 cells expressing the P2X7 wild-type or the 489T mutant receptor was 132 ± 20 and 196 ± 22 nM (mean ± SEM, from five separate experiments; p < 0.01 by Student’s t test for paired data), respectively. Analysis of P2X7 489T mutant receptor permeability was extended to larger cations such as ethidium. Recombinant HEK293 cells were incubated in the presence of extracellular ethidium bromide and, after reaching a stable baseline, exposed to ATP, and the rate of ethidium uptake was measured. Response to extracellular ATP was calculated as percent of permeability increase caused by digitonin, as described in Materials and Methods. In clones with similar surface expression levels of the P2X7R (Fig. 5B), the rate of ethidium uptake was higher in the P2X7 489T than in the wild-type HEK293 cells, as shown in the representative time course of Fig. 5A. In these experiments, the P2X7 1513C HEK293 cell line showed a very low ATP-dependent ethidium uptake (Fig. 5A). The average percent permeability increase of the P2X7 wild-type HEK293 cells was 30.1 ± 0.3, whereas that of P2X7 489T HEK293 cells was 44.4 ± 4.4 (mean ± SEM, from four separate experiments; p < 0.05 by Student’s t test for paired data). Therefore, the functional analyses of P2X7R mutants in vitro indicate that the P2X7 489T allele confers a significant gain of receptor function, that the 1068A variant causes a slight functional reduction, and that the 1096G and 1405G variants have an almost neutral effect.
Discussion
The main finding of this functional study in ex vivo lymphocytes from CLL patients and in in vitro recombinant mutant cell lines is that the 489C>T variant of the P2X7 gene is associated with gain-of-function of the receptor, that the 1068G>A variant slightly reduces receptor activation, and that the 1096C>G and 1405A>G variants have an almost neutral effect. In addition, we confirm that the 1513A>C polymorphism causes loss-of-function of the receptor, and we report that the prevalence of these five variants of the P2X7 gene is similar in patients affected by CLL and healthy control individuals.
A dual role of the P2X7R is widely accepted (24), because it promotes cell apoptosis upon prolonged exposure to high ATP concentrations, whereas at tonic, low concentrations of the ligand, it induces cell growth (7). The modulation of P2X7R function is attracting interest for its potential implications in human pathology, and in particular, inheritance of a defective P2X7R has been claimed to be associated to predisposition to CLL. In a study conducted in an Australian sample population, the prevalence of the P2X7 1513A>C loss-of-function polymorphism was 3-fold greater in patients with CLL than in controls (20). Wiley et al. suggested that inheritance of the 1513C allele could contribute to the pathogenesis of CLL, as in the presence of this allele, P2X7-dependent apoptosis is impaired, and this could result in accumulation of neoplastic B cells. Accordingly, they proposed P2X7 as a tumor suppressor gene. However, subsequent studies performed in CLL patients from Sweden, the United Kingdom, Germany (28); extension of the original paper in familial cases (21, 23, 29, 30); and the present study create a consensus on the observation that the P2X7 1513C allele prevalence is not increased in the sporadic form of CLL, which makes less likely the direct involvement of the P2X7 1513A>C polymorphism in the pathogenesis of CLL.
Truncation or partial deletions of the P2X7 protein in the C-terminal protein tail impair the ATP-dependent transmembrane ion transport, indicating that different domains within the C-terminal end of the receptor either facilitate the assembly of the monomers into the channel complex or are relevant to target the receptor in the plasma membrane (1, 2, 9). Although extended deletions substantially modify P2X7R activation, single residue changes within the same domain have unpredictable effects. The 1513A>C sequence variant, encoding the Glu-496 to Ala change, was found to impair both Ba2+ and ethidium uptake in blood peripheral mononuclear and transiently transfected recombinant HEK293 cells (10). By studying P2X7R-dependent intracellular Ca2+ rise, we confirm that the 1513C allele results in loss-of-function of the receptor both in lymphocytes of CLL patients ex vivo and in recombinant HEK293 cells in vitro (Figs. 1 and 3). Our data do not provide insights on the mechanism responsible for this functional defect, albeit it might be speculated that the Glu to Ala residue change might weaken the electrostatic interactions during the assembly of the P2X7 channel complex in the plasma membrane (10). Here we investigated three additional allelic variants changing the sequence of the C-terminal protein tail, namely the amino acid changes Ala-348 to Thr (1068G>A), Thr-357 to Ser (1096C>G), and Gln-460 to Arg (1405A>G), which are in proximity to the Glu-496 to Ala variation. The functional analysis of these amino acid changes conducted both in lymphocytes from CLL patients and in P2X7-transfected HEK293 cells indicate a slight reduction (1068A) or neutral effect (1096G and 1405G) on receptor activation (Figs. 2 and 3), although the possibility remains open that combined residue changes in the P2X7R could alter channel activation more profoundly.
Little is known about the pocket allowing ATP binding to the extracellular loop of the receptor. Positively charged residues, which are candidate binding sites for ATP, have been mutated and expressed in recombinant cell models, and both Lys-193 and Lys-311 were shown to be essential in ATP binding to the extracellular loop of the human P2X7R (31). More recently, the single nucleotide substitution 946G>A, encoding the Arg-307 to Gln change, has been identified in Caucasian subjects with an allelic frequency of 0.014 (32). Interestingly, P2X7R-dependent Ba2+, Rb+, and ethidium+ influx was markedly reduced in PBL and monocytes obtained from 946G/A heterozygous individuals, whereas P2X7R cell surface expression was conserved, indicating that Arg-307 is critical for ATP binding and receptor activation (32). In this study, we found that the frequent 489C>T allelic variant encoding the change of His-155 to Tyr is associated with an increased P2X7R-dependent Ca2+ influx in transfected HEK293 cells (Fig. 4). Interestingly, gain of function associated with the 489T allele was also observed in PBL from 1513A/A homozygous CLL patients (Fig. 1) but not in 1513A/C and 1513C/C subjects, indicating that the loss- of-function conferred by the E496A residue change in the long C-terminal tail cannot be recovered by the H155Y substitution in the extracellular loop of the receptor. As the list of P2X7 allelic variants conferring loss of receptor function is not thoroughly defined, it could be possible that the combined presence of other unknown loss-of-function alterations on the same copy of the P2X7 gene might reduce or abolish the gain of function conferred by the 489T allele in blood PBL ex vivo. Thus, the finding of the 489T allele in CLL individuals is expected to increase P2X7R function in peripheral blood lymphocytes to variable extents, depending on the combined presence of loss-of-function variants. How might the change of His-155 to Tyr in the extracellular loop of the receptor increase the P2X7R channel function? Considering the high frequency of the 489T minor allele that we observed in CLL patients and healthy control individuals, a large fraction of the human P2X7 protein should have a Tyr residue in position 155 of the primary sequence. Extensive molecular studies have been performed on the role of 19 Tyr residues of rat P2X7R, both on activation and on the association of P2X7R with accessory proteins, as a function of the phosphorylated status (33), whereas studies on the residue homologous to the H115Y variation of the human sequence protein are missing. The most straightforward explanation might take into consideration that the H155Y change increases the affinity of the agonist to the receptor, although the ATP and Bz-ATP used in our experimental conditions were at saturating concentrations. Considering the change of electrostatic charges consequent to an His to Tyr substitution and the potential of Tyr phosphorylation, alternative speculations could include a facilitation in the recruitment of the P2X7 monomers in the process of assembly of the receptor complexes or a stabilization of the active complexes themselves, thus increasing the ratio between active complexes and inactive monomers. Additional experiments are needed to fully clarify this issue.
Acknowledgments
We thank Gary Buell (Serono Pharmaceutical Research Institute, Geneva, Switzerland), for providing the hP2X7p635 plasmid and the anti-P2X7R mAb, and Richard Rosenquist (Rudbeck Laboratory, University of Uppsala, Uppsala, Sweden) for thoughtful discussions.
Disclosures
The authors have no financial conflict of interest.
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 Ministry of Education and Scientific Research (grants Cofin and Firb), Italian Association for Cancer Research (AIRC) and institutional funds from the University of Ferrara.
2 Address correspondence and reprint requests to Dr. Giulio Cabrini, Laboratorio Patologia Molecolare, Centro Fibrosi Cistica, Azienda Ospedaliera Verona, Piazzale Stefani 1, 37126 Verona, Italy. E-mail address: giulio.cabrini@azosp.vr.it
3 Abbreviations used in this paper: Bz-ATP, 2',3'-O-(benzoyl-4-benzoyl)-ATP; HEK, human embryonic kidney; TSC, The Single Nucleotide Polymorphism Consortium; CLL, chronic lymphocytic leukemia; [Ca2+]i, intracellular Ca2+ concentration; UNG, uracil-N-glycosylase; MGB, minor groove binder; SNP, single nucleotide polymorphism.
Received for publication December 23, 2004. Accepted for publication April 12, 2005.
References
Surprenant, A., F. Rassendren, E. Kawashima, R. A. North, G. Buell. 1996. The cytolytic P2Z receptor for extracellular ATP identified as a P2X receptor (P2X7). Science 272: 735-738.
North, R. A.. 2002. Molecular physiology of P2X receptors. Physiol. Rev. 82: 1013-1067.
Rassendren, F., G. N. Buell, C. Virginio, G. Collo, R. A. North, A. Surprenant. 1997. The permeabilizing ATP receptor, P2X7, cloning and expression of a human cDNA. J. Biol. Chem. 272: 5482-5486
Buell, G., F. Talabot, A. Gos, J. Lorenz, E. Lai, M. A. Morris, S. E. Antonorakis. 1998. Gene structure and chromosomal localization of the human P2X7 receptor. Receptor Channel 5: 347-354.
Nicke, A., H. G. B?umert, J. Rettinger, A. Eichele, G. Lambrecht, E. Mutschler, G. Schmalzing. 1998. P2X1 and P2X3 receptors form stable trimers: a novel structural motif of ligand-gated ion channels. EMBO J. 17: 3016-3028.
Virginio, C., A. MacKenzie, R. A. North, A. Surprenant. 1999. Kinetics of cell lysis, dye uptake and permeability changes in cells expressing the rat P2X7 receptor. J. Physiol. 519: 335-346.
Baricordi, O. R., L. Melchiorri, E. Adinolfi, S. Falzoni, P. Chiozzi, G. Buell, F. Di Virgilio. 1999. Increased proliferation rate of lymphoid cells transfected with the P2X7 ATP receptor. J. Biol. Chem. 274: 33206-33208.
Nuttle, L. C., G. R. Dubyak. 1994. Differential activation of cation channels and nonselective pores by macrophage P2Z purinergic receptors expressed in Xenopus oocytes. J. Biol. Chem. 269: 13988-13999.
Smart, M. L., B. J. Gu, R.G. Panchal, J. S. Wiley, B. Cromer, D. A. Williams, S. Petrou. 2003. P2X7 receptor cell surface expression and cytolytic pore formation are regulated by a distal C-terminal region. J. Biol. Chem. 278: 8853-8860.(Giulio Cabrini2,*, Simone)