Neural progenitor cell death is induced by extracellular ATP via ligation of P2X7 receptor

Authors


Address correspondence and reprint requests to Jean M. Kanellopoulos, Laboratoire Activation Cellulaire et Transduction des Signaux, Institut de Biochimie et Biophysique Moléculaire et Cellulaire, Univ Paris-Sud, CNRS, UMR 8619, Bâtiment 430, 91405 Orsay Cedex, France. E-mail: jean.kanellopoulos@u-psud.fr

Abstract

J. Neurochem. (2009) 109, 846–857.

Abstract

Neural progenitor cells (NPCs) are capable of self-renewal and differentiation into neurons, astrocytes and oligodendrocytes, and have been used to treat several animal models of CNS disorders. In the present study, we show that the P2X7 purinergic receptor (P2X7R) is present on NPCs. In NPCs, P2X7R activation by the agonists extracellular ATP or benzoyl ATP triggers opening of a non-selective cationic channel. Prolonged activation of P2X7R with these nucleotides leads to caspase independent death of NPCs. P2X7R ligation induces NPC lysis/necrosis demonstrated by cell membrane disruption accompanied with loss of mitochondrial membrane potential. In most cells that express P2X7R, sustained stimulation with ATP leads to the formation of a non-selective pore allowing the entry of solutes up to 900 Da, which are reportedly involved in P2X7R-mediated cell lysis. Surprisingly, activation of P2X7R in NPCs causes cell death in the absence of pore formation. Our data support the notion that high levels of extracellular ATP in inflammatory CNS lesions may delay the successful graft of NPCs used to replace cells and repair CNS damage.

Abbreviations used
BAPTA-AM

1,2-bis-(o-aminophenoxy)-ethane-N,N,-N′,N′-tetraacetic acid tetraacetoxy-methyl ester

Bz-ATP

benzoylbenzoyl ATP

DMEM

Dulbecco’s modified Eagle’s medium

DR

death receptors

DRM

detergent-resistant membrane

EAE

experimental autoimmune encephalomyelitis

EGF

epidermal growth factor

Erk

extracellular-signal regulated kinase

EtBr

ethidium bromide

FCS

foetal calf serum

JNK

c-jun N-terminal kinase

LDH

lactate dehydrogenase

MTT

3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide

NPC

neural progenitor cell

o-ATP

oxidized ATP

P2X7R

P2X7 receptor

TBS

Tris-buffered saline

TMRE

tetramethylrhodamine ethyl ester perchlorate

Neural progenitor cells (NPCs) constitute a heterogeneous cell population capable of self-renewal and differentiation into neurons, astrocytes and oligodendrocytes. In the adult rodent, these cells are restricted to two regions of the CNS: the subventricular zone of the lateral ventricles (Alvarez-Buylla and Garcia-Verdugo 2002) and the dentate gyrus of the hippocampus (Gage et al. 1998). In embryonic mice, the majority of neuroepithelial cells have progenitor cell properties (Kilpatrick and Bartlett 1993). NPCs can be isolated and propagated as neurospheres, in the presence of epidermal growth factor (EGF) and fibroblast growth factor (Reynolds and Weiss 1996).

Neural progenitor cell-based therapy has been successfully developed for animal models of CNS disorders such as Parkinson’s disease, Huntington’s disease, stroke injury and multiple sclerosis (Martino and Pluchino 2006). Grafting NPCs during experimental autoimmune encephalomyelitis (EAE), an animal model of multiple sclerosis, improves remyelination and induces a decrease in T-cell infiltrates (Martino and Pluchino 2006). Furthermore, endogenous NPCs are recruited during EAE and migrate into areas of demyelination, where they differentiate into glial cells (Picard-Riera et al. 2002).

However, during neurodegenerative and demyelinating diseases and stroke, tissue damage and inflammation lead to the release of various cytokines and mediators as well as high levels of extracellular pro-inflammatory nucleotides such as ATP (Fields and Burnstock 2006; Khakh and North 2006). This increase in local concentrations of ATP could enhance inflammation mediated by purinergic receptors and/or lead to cell death.

P2X7 receptor (P2X7R) is a peculiar purinergic receptor as it acts both as a classical ATP-gated ion channel and can also induce cell death. This receptor is a transmembrane protein whose ligation of ATP4− leads to opening of cation channels, in particular an increase in intracellular Ca2+ concentrations (Virginio et al. 1999; North 2002). Prolonged exposure to ATP induces the formation of a non-selective pore allowing the entry of solutes up to 900 Da, which eventually leads to cell death (Virginio et al. 1999; North 2002).

Cell death has been observed in different cells of haematopoietic origin such as lymphocytes (Ferrari et al. 1999), thymocytes (Zheng et al. 1991; Auger et al. 2005), macrophages and dendritic cells (Coutinho-Silva et al. 1999) as well as in neuronal cells such as the dopaminergic neuronal immortalized cell line SN4741 (Jun et al. 2007) and retinal cholinergic neurons (Resta et al. 2005). In addition to its role in cell death, the P2X7R is involved in release of inflammatory cytokines, such as IL-1β, IL-6 and tumour necrosis factor-α (Labasi et al. 2002; Ferrari et al. 2006; Fields and Burnstock 2006). Studies on the functional role of P2X7R in the nervous system are actually an extensive area of research. P2X7R-mediates neuromediator release, like glutamate, endocannabinoids and ATP, from astrocytes (Duan et al. 2003; Walter et al. 2004; Suadicani et al. 2006), as well as from neurons in the hippocampus, cortex and spinal cord and motor neurons (Deuchars et al. 2001; Sperlagh et al. 2006; Marcoli et al. 2008).

These findings raise the possibility that the activation of P2X7R may contribute to cell death in CNS disorders. Effectively, administration of P2X7R antagonists in rat acute spinal cord injuries improves functional recovery and decrease cell death in the periphery of traumatic lesions (Wang et al. 2004). Matute et al. (2007) have also demonstrated that during chronic EAE, ATP can induce oligodendrocyte cell death via P2X7R activation and contribute to the demyelination process.

Therefore, the therapeutic potential of NPCs could also be affected negatively by P2X7R stimulation in CNS disorders. In this study, we have investigated the functional properties of P2X7R in NPCs and in particular, the sensitivity to P2X7R-mediated NPC death as these cells represent a potential cell source for cell replacement therapy.

Materials and methods

Animals

Four- to 8-week-old C57BL/6 mice were purchased from Charles River France (I'Arbresle, France). P2X7R-deficient mice were obtained from Pfizer Inc. (Groton, CT, USA). Mice used in this work were from fourth backcross onto C57BL/6 (Labasi et al. 2002).

Reagents and antibodies

Fibroblast growth factor was obtained from PeproTech France (Neuilly-Sur-Seine, France), B27 supplement and Dulbecco’s modified Eagle’s medium (DMEM)/F12 medium from GIBCO Life Technologies (Cergy Pontoise, France). EGF, insulin, ATP, benzoylbenzoyl ATP (Bz-ATP), oxidized ATP (o-ATP), EGTA, lactate dehydrogenase (LDH) detection kits and poly-ornithine were purchased from Sigma (St Louis, MO, USA). All pharmacological signalling pathway inhibitors were obtained from Calbiochem (San Diego, CA, USA): U0126, SB203580, SP600125, anisomycin, Z-VAD-FMK and 1,2-bis-(o-aminophenoxy)-ethane-N,N,-N′,N′-tetraacetic acid tetraacetoxy-methyl ester (BAPTA-AM). Ethidium bromide (EtBr) was obtained from Invitrogen (Carlsbad, CA, USA).

As primary antibodies, we used the following: rabbit anti-glial fibrillary acidic protein (1 : 200; Dako, Ely, UK), monoclonal anti β-III-tubulin (1 : 200; Sigma), monoclonal anti-nestin (1 : 200; Chemicon, Temecula, CA, USA), monoclonal anti-O4 (hybridoma supernatant diluted 1 : 10). goat anti-rabbit IgG (Dako), rat anti-mouse IgG (P.A.R.I.S), goat anti-mouse IgM (Sigma) were used as secondary antibodies at 1 : 100.

The following antibodies were used for western blotting: unconjugated rabbit anti-extracellular-signal regulated kinase (Erk) 1/2 (pTpY185/187), anti-c-jun N-terminal kinase (JNK) 1/2 SAPK (pTpY183/185), anti-p38 (pTpY180/182) phospho-specific antibodies (Biosource International, Camarillo, CA, USA), rabbit anti-Erk1/2 (Zymed Laboratories, South San Francisco, CA, USA), anti-p38 (C20) and goat anti-actin (I-19), affinity purified rabbit anti-carboxyterminal region of rat P2X7R (Alomone Labs, Jerusalem, Israel). Affinity purified goat anti-rabbit IgG antibodies coupled to peroxidase (Rockland Immunochemicals, Gilbertsville, PA, USA), goat anti-mouse IgG antibodies coupled to peroxidase, mouse monoclonal anti-goat/sheep IgG antibodies conjugated to peroxidase (Sigma-Aldrich), affinity purified donkey anti-rabbit IgG antibodies coupled to IRDye 800 (Rockland), affinity purified rabbit anti-goat IgG antibodies coupled to Alexa Fluor 680 (Invitrogen) were used as secondary antibodies for western blot analyses.

Cell cultures

Neural progenitor cell cultures were prepared from striatum of mice at E14.5 day. Briefly, spheres of NPCs were grown in DMEM/F12 supplemented with B27, EGF (20 ng/mL), fibroblast growth factor (10 ng/mL) and insulin (20 μg/mL) (NPC medium) (Reynolds et al. 1992). Two-thirds of the culture medium was changed every 2–3 days. Neuro-2a neuroblastoma cells were maintained in DMEM containing 10% fetal calf serum (FCS). Primary cultures of astrocytes were generated from hemispheres of 1- to 3-day-old mice. Cells were grown in culture medium containing DMEM, glutamax, non-essential amino acids, sodium pyruvate, penicillin–streptomycin and 20% FCS. The medium was changed twice weekly. After 1 week, the concentration of FCS was reduced to 10%. Macrophages were obtained from C57BL/6 mice, 5 days after intraperitoneal injection of thioglycollate.

RT-PCR

Briefly, 2 μg of total RNA, extracted from cells using Trizol reagent (Invitrogen), were reverse transcribed using SuperScript First-Strand Synthesis system for RT-PCR (Invitrogen). PCR reactions were performed using TaKaRa LA Taq (Lonza) in a Gene Amp PCR system 9700 (Applied Biosystems, Foster City, CA, USA). The following mouse primers were used: P2X7R forward 5′-TTCCAGGAAGCAGGAGAGAA-3′, P2X7R reverse 5′-ATACTTCAACGTCGGCTTGG-3′, Pannexin-1 forward 5′-GAGTCTGTGGGAGATATCTG-3′, Pannexin-1 reverse 5′-TGCAGCCTCGCTGCTCAGG-3′.

Electrophysiology

Electrophysiological experiments were performed with neurospheres seeded on poly-ornithine-coated 12 mm round coverslips and cultured for 1–2 days before recordings. Patch-clamp voltage-clamp experiments (whole-cell configuration) were performed at 20°C with cultures placed in the recording chamber and continuously superfused (2 mL/min) with the extracellular solution. Cells were visually identified using an upright microscope (Axioskop; Carl Zeiss France, Le Pecq, France) equipped with differential interference contrast optics and a water immersion 63X, 0.9 numerical aperture objective. The extracellular solution contained the following (in mM): 156 NaCl, 4 KCl, 2 CaCl2, 10 HEPES and 10 glucose. In some experiments 1 mM MgSO4 was added to the solution. The pH was adjusted to 7.4 using NaOH. This solution contained 1–2 μM 8-cyclopentyl-1,3-dipropylxanthine (Tocris Bioscience, Ellisville, MI, USA), an antagonist of A1 adenosine receptors. Bz-ATP was bath applied. The internal solution contained the following (in mM): 140 KGlu, 6 KCl, 10 HEPES, 0.2 EGTA, 1 MgCl2, 0.4 Na-GTP, 4 Na2-ATP, pH 7.3 with KOH; the osmolarity was 295–300 mOsm/l. Pipettes were pulled from thick borosilicate glass capillaries and had a resistance of 7–10 MΩ when filled with intracellular solutions. Cells were voltage clamped at −60 mV and no correction for liquid junction potential was performed. Membrane currents were recorded using an Axopatch 200 amplifier (Axon Instruments, Union City, CA, USA). They were filtered at 2 kHz and sampled at 5 kHz with a 1322A Digidata interface (Axon Instruments). Series resistance was not compensated but was monitored throughout the experiment using a −10 mV step. Acquisitions were performed with Elphy software and off-line analyses were performed using the Clampfit program (Axon Instruments). Averages were expressed as mean ± SEM. The Mann–Whitney test was used to identify statistical significance among data groups.

MTT assay

Cell viability was monitored by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. Cells were seeded in 96-well plates (50 000 cells/well) and then treated with various concentrations of ATP or Bz-ATP in triplicate. Two hours before the end of the experiment, MTT reagent (5 mg/mL) was added to each well at the final concentration of 1 mg/mL. Lysis buffer (125 μL/well) was added to stop the reaction. The plate was read for optical density at 540 nm using a spectrofluorometer (Wallac 1420, PerkinElmer France, Courtaboeuf, France).

Assay for mitochondrial membrane potential

Cells were incubated for 30 min in DMEM/F12 with JC-1 dye (Sigma) at a final concentration of 4.8 μM and treated with 2.5 mM Bz-ATP. At the end of the experiment the cells were washed and analysed by flow cytometry (Becton-Dickinson, Franklin Lakes, NJ, USA).

Another method described by Voronina et al. (2004) was used to assess mitochondrial membrane depolarisation. In this technique, large amounts of the fluorescent probe tetramethylrhodamine ethyl ester perchlorate (TMRE) (Molecular probes, Invitrogen) accumulate into mitochondria. At high concentrations of TMRE, the fluorescence is quenched. However, when mitochondrial membrane is depolarized, TMRE is released from the mitochondria and an increase in fluorescence is measured because of fluorescent probe ‘dequenching’. Cells were loaded with 1 μM TMRE in DMEM/F12 supplemented with B27 (1 mM Ca2+) for 30 min at 37°C. Then, cells were washed in modified Krebs-HEPES medium and suspended in this buffer. The modified Krebs-HEPES contained: 128 mM NaCl, 2.5 mM KCl, 2.7 mM CaCl2, 16 mM glucose and 20 mM HEPES. The pH was adjusted to 7.4 using NaOH. Fluorescence was monitored for 30 min at 37°C in a spectrofluorometer using an excitation wavelength of 550 nm and an emission wavelength of 590 nm and Bz-ATP was added after 50 s and applied for 200 s.

Measurement of LDH release

Unless otherwise specified in the figure legends, NPCs were pre-incubated in DMEM/F12 for 1 h with the pharmacological inhibitors before addition of Bz-ATP for 3 h. Cell lysis was quantified by measuring the release of LDH. Cells (1 × 105) were incubated in 100 μL of DMEM/F12 for various lengths of time. Cells were centrifuged at 200 g for 10 min, and supernatants were tested for LDH release using the oxidation reaction of β-NADH in the presence of pyruvate (Sigma kit for LDH). The initial rate of absorbance decrease was measured in an automatic readout spectrophotometer at λ = 340 nm.

BAPTA-AM and EGTA experiments

Neural progenitor cells were incubated in DMEM/F12 medium supplemented containing B27 (Ca2+ concentration = 1 mM), with or without 5 mM EGTA to chelate extracellular Ca2+ and/or 50 μM BAPTA-AM to prevent intracellular calcium increase. After 1 h at 37°C, Bz-ATP is added directly to the cell suspensions for 3 h at 37°C and percentage of LDH release is determined.

Calcium experiments

Cells were pre-incubated with or without 50 μM BAPTA-AM for 1 h and loaded with 2 μM fura-2AM (Molecular probes) in DMEM/F12 supplemented with B27 (1 mM Ca2+) for the last 30 min at 37°C. Then, cells were washed in modified Krebs-HEPES medium and suspended in this buffer. The modified Krebs-HEPES contained: 128 mM NaCl, 2.5 mM KCl, 2.7 mM CaCl2, 16 mM glucose, 20 mM HEPES, with or without 5 mM EGTA. The pH was adjusted to 7.4 using NaOH. We measured the increase in Ca2+ in NPCs by dual excitation spectrofluorimetric analysis at 340 and 380 nm (ratio of OD340/OD380) and Bz-ATP was added after 50 s.

Western blot analyses

Lysates or immunoprecipitates from cells were analysed by sodium dodecyl sulphate–polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes, which were blocked with 4% non-fat-milk in Tris-buffered saline (TBS) containing 0.2% Tween-20 for 1 h at 37°C. Blots were immunostained with primary antibodies at 4°C overnight and probed with secondary antibodies conjugated to horseradish peroxidase. Specific bands were visualized by enhanced chemiluminescence (PerkinElmer France).

Non-selective pore formation

For EtBr uptake experiments, 0.5 × 106 cells were seeded in poly-ornithine-coated flat bottom 96-well plates and cultured for 24 h at 37°C. Cells were incubated with 10 μg/mL of EtBr diluted in solution containing 1 mM Ca2+ for 10 min. Bz-ATP was added and fluorescence was monitored for 20 min at 37°C in a spectrofluorometer (Wallac 1420) using an excitation wavelength of 485 nm and an emission wavelength of 615 nm.

Detergent-resistant membrane preparation

Cells were seeded in poly-ornithine-coated flask and cultured at 37°C. Cells were submitted to surface biotinylation using sulfo-NHS-LC-Biotin (Pierce, Rockford, IL, USA) at 1 mg/mL at 4°C for 30 min. Five million cells were resuspended in TBS [Tris 50 mM, NaCl 150 mM, EDTA 1 mM (pH 7.4)] containing 0.05% Triton X-100 and protease inhibitors (complete tablets; Roche, Basel, Switzerland) and passed six times through a 26G3/8 needle. After 20 min incubation on ice, lysates were adjusted to 45% sucrose and placed at the bottom of a SW41 centrifuge tube. A sucrose step gradient was performed by layering 6 mL of 36% and 3.5 mL of 5% (w/v) sucrose in TBS. Sucrose percentages were assessed by refractometry. After centrifugation at 267 000 g (at the bottom of the tube) at 4°C for 15 h, 2 mL aliquots of the gradients were collected from the top and submitted to precipitation with neutravidin-agarose beads (Pierce) at 4°C overnight. Neutravidin-bound proteins from each sucrose gradient fraction were separated by sodium dodecyl sulphate–polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes. Blots were immunostained with the affinity purified rabbit polyclonal anti-P2X7R antibodies.

Statistical analysis

Data are expressed as a mean value ± SD. Data were analysed by using Student’s t-test and p < 0.05 (at least) was considered statistically significant.

Results

ATP induces NPC death

We first determined what was the effect of sustained stimulation of NPC with ATP. We observed that ATP induces cell death of NPCs, as measured by the MTT assay (Fig. 1a) with concentration of 3 mM ATP. NPC cell death is detected as early as 2 h following stimulation with 5 mM ATP (Fig. 1b). The high concentrations of ATP needed to induce cell death suggested the involvement of the P2X7R.

Figure 1.

 ATP stimulation induces NPC cell death via activation of P2X7R. Panel (a), MTT test was performed on NPCs treated with increasing concentrations of ATP during 3 h at 37°C. Panel (b), MTT test was performed on NPCs treated with 5 mM ATP for increasing lengths of time. Panel (c), NPCs and brain from C57BL/6 mice were isolated and RNA was extracted as described in Materials and methods. Two micrograms of total RNA were reverse transcribed, and P2X7R expression was determined by RT-PCR (c, left panel). Immunoblots were performed with affinity purified rabbit anti-P2X7R antibodies on NPCs lysates of wt and P2X7Rko mice (c, right panel). The same blots were stripped and probed for actin to show equal loading (c, intermediate panel). NPCs were surface-labelled with sulfo-NHS-LC-biotin and lysed in 0.05% Triton X-100 before fractionation on a 36%/5% sucrose density gradient. Neutravidin-bound proteins from fraction 1 to 5 were resolved on SDS–PAGE and analysed by western blot/ECL to detect P2X7R. The low-density fraction (fraction 2) contains DRMs. The heavy density fractions (fraction 45) contain the bulk of solubilised membrane proteins (c, lowest panel). Panel (d), representative whole-cell patch clamp recordings of NPC from wild-type mice (top) and P2X7Rko mice (bottom) in the absence of extracellular Mg2+. Top: bath application of 100 μM Bz-ATP induced an inward current that slowly increased, reached a peak of −90 pA and then partially returned to baseline (dashed line) during Bz-ATP washout. Bottom: in a P2X7Rko cultured cell Bz-ATP failed to elicit any current change. Panel (e), MTT test was performed on NPCs treated with increasing concentrations of Bz-ATP during 3 h at 37°C. Panel (f), NPCs were pre-incubated for 2 h at 37°C with 0.5 mM of o-ATP, a specific inhibitor of P2X7R. NPCs were then treated with 5 mM ATP for 3 h. Panel (g), MTT test was performed on NPCs of wt and P2X7Rko mice treated with 5 mM ATP during 3 h at 37°C. Panel (h), MTT test was performed on NPCs of wt and P2X7Rko mice treated with 2.5 mM Bz-ATP during 3 h at 37°C. The experiments were performed on separate days, and the values represent the mean and SD of three experiments. **p < 0.01 and ***p < 0.001.

NPCs express functional P2X7R

We next determined that NPCs from wild type C57BL/6 mice express the P2X7R transcripts as well as the P2X7R protein, as detected, respectively, by RT-PCR and western blot analysis (Fig. 1c). No band corresponding to the P2X7R protein was observed in NPCs from P2X7R-deficient mice (Fig. 1c). It has recently been shown that a fraction of P2X7R present at the plasma membrane is associated with detergent-resistant membranes (DRMs) (Garcia-Marcos et al. 2006; Barth et al. 2007) and that this distribution is due to palmitoylation of P2X7R (Gonnord et al. 2009). Thus, we determined whether P2X7R is expressed at the plasma membrane of NPCs and is equally distributed between low- and high-density sucrose fractions as observed in different cell types (Garcia-Marcos et al. 2006; Barth et al. 2007; Gonnord et al. 2009). NPCs were surface labelled with biotin, lysed in 0.05% Triton X-100 and fractionated on sucrose gradient. Biotinylated proteins were isolated on neutravidin beads and analysed by western blot. As shown in Fig. 1(c) P2X7R is found at the plasma membrane and similar amounts of P2X7R are found in DRMs and non-DRM fractions as previously described (Garcia-Marcos et al. 2006; Barth et al. 2007; Gonnord et al. 2009).

Electrophysiological experiments were then performed to ensure that the P2X7R of NPCs is functional. In NPCs, by RT-PCR we found the transcripts of the seven P2XR (not shown). Thus, we used Bz-ATP for patch-clamp experiments as it is the most potent agonist at this subtype (North 2002). Patch-clamp voltage-clamp recordings onto NPCs revealed a resting membrane potential that ranged between −54 and −75 mV (mean value: −63.5 ± 3.6 mV, = 13). The functional expression of P2X7R in cultured NPC was examined in response to bath application of Bz-ATP (100 μM) (Fig. 1d). In cultures from wild-type mice, in a normal divalent cation containing solution, Bz-ATP evoked inward currents of small amplitudes (8.7 ± 2.1 pA, = 3, not shown). Therefore all following experiments were carried out in the absence of extracellular Mg2+. This favours P2X7R opening and potentiates the response to Bz-ATP (North 2002). Under these conditions, Bz-ATP induced an inward current (see example in Fig. 1d, top) that reached a mean peak amplitude value of −59.4 ± 12.9 pA (= 6) in 50% of recorded cells (six out of 12 cells). As shown in Fig. 1(d) (top), the baseline current was partially recovered after wash out with control solution. Identical experimental protocols were carried out with NPC cultures from P2X7R-deficient mice. Bz-ATP could not induce any inward currents (−1.3 ± 0.5 pA = 7) in any of the P2X7R-deficient cells, as shown in Fig. 1(d) (bottom). These data further support membrane expression of P2X7R in NPCs and that P2X7R are functional in a population of NPCs from wild-type mice.

Activation of P2X7R induces NPC death

The involvement of P2X7R in NPC death is suggested by the greater potency of Bz-ATP compared to ATP; indeed NPC cell death was observed with concentration of 3 mM ATP and 0.3 mM Bz-ATP (Fig. 1a and e) (North 2002). Moreover, pre-treatment of NPCs with o-ATP, a known inhibitor of P2X7R, prevented ATP-induced death of NPCs (Fig. 1f). Importantly, addition of extracellular ATP or Bz-ATP leads to cell death of NPCs from C57BL/6 mice but not of NPCs derived from P2X7R-deficient animals (Fig. 1g and h). Taken together, these results indicate that NPC cell death is due to ATP- or Bz-ATP-specific action onto P2X7R.

Neuronal and glial progenitor cells are sensible to P2X7R-induced cell death

Neurospheres contain a small percentage (1–5%) of slowly proliferating neural stem cells and a majority of rapidly growing nestin-positive precursor cells arising from a single stem cell (Reynolds and Weiss 1996). No specific markers are available to unequivocally identify neuronal or glial progenitors. Indeed, their ability to differentiate into neurons, astrocytes and oligodendrocytes provide the main features for their identification. Thus, we determined whether P2X7R stimulation selects for a sub-population of death-resistant cells among NPCs that is already committed to a differentiation pathway (neuronal versus glial progenitors). Our data demonstrate that death-resistant cells maintain their differentiation potential and are able to generate the same proportion of oligodendrocytes (O4 labelled cells), astrocytes (glial fibrillary acidic protein labelled cells) and neurons (β-III-tubulin labelled cells) as untreated NPCs (Fig. 2).

Figure 2.

 P2X7R activation does not select a sub-population of death-resistant cells. Panel (a), after treatment or not with 2.5 mM Bz-ATP during 4 h, NPCs were cultured on poly-ornithine substrate with 1% FCS to induce cell differentiation, or without FCS to maintain them as neural precursor cells. 5 days after adhesion, cells were fixed and labelled with antibodies against nestin (neural precursor), β-III-tubulin (neuron), GFAP (astrocyte) and O4 (oligodendrocyte). Panel (b), the percentage of stained cells was established by counting the number of positive cells for each antibody. The experiments were performed on separate days, and the values represent the mean and SD of three experiments.

Ligation of P2X7R in NPCs leads to cell death by lysis/necrosis

P2X7 receptor can be involved in both apoptosis and/or necrosis depending on the cell type (Di Virgilio et al. 1998; Ferrari et al. 1999; Auger et al. 2005). In order to determine if P2X7R induce NPC apoptosis, we used the pan caspase inhibitor, Z-VAD-FMK. As can be seen in Fig. 3a, Z-VAD-FMK blocks staurosporine induced cell death while it does not protect NPCs from ATP-induced cell death (Fig. 3a), ruling out a role of caspases in this process.

Figure 3.

 P2X7R activation leads to necrosis rather than apoptosis of NPCs. Panel (a), NPCs were pre-incubated for 1 h at 37°C with 50 μM of Z-VAD-FMK inhibitor, a broad inhibitor of caspase. NPCs were then treated with 5 mM ATP for 3 h (left panel) or with 250 nM staurosporine for 24 h (right panel). Panel (b), kinetic studies of LDH release from NPCs of wt and P2X7Rko mice treated with 2.5 mM Bz-ATP for 5 h. Panel (c), kinetics of LDH release from NPCs treated with 2.5 mM Bz-ATP for 1 h. Specific LDH release was calculated by subtracting LDH release values of unstimulated cells from values of Bz-ATP-stimulated cells. For panels a–c, the experiments were performed on separate days, and the values represent the mean and SD of three experiments. **p < 0.01.

Neural progenitor cell plasma membrane disruption was assessed by measuring the release of lactate dehydrogenase (LDH), a cytosolic protein. As shown in Fig. 3b, LDH release, reaches a plateau 3 h following the addition of 2.5 mM Bz-ATP, with a maximal LDH release of 40%. In contrast, LDH release was only 5% in NPCs derived from P2X7R-deficient animals. Furthermore, loss of plasma membrane integrity is significantly detected 30 min after treatment with 2.5 mM Bz-ATP (Fig. 3c). Taken together, these findings show that stimulation of P2X7R induces NPC cell death by necrosis.

MAPK modules are not triggered by P2X7R activation in NPCs

The signalling pattern leading to death by necrosis is less defined than in cell death by apoptosis. Thus, we first wanted to verify if the pattern of activation of MAPKs involved in ATP-induced necrosis of thymocytes could apply to NPCs (Auger et al. 2005).

The activation of the three MAPK modules following the ligation of thymocyte P2X7R by ATP has been reported by our laboratory (Auger et al. 2005). In contrast, as illustrated in Fig. 4(a), Bz-ATP stimulation of NPCs does not lead to the phosphorylation of Erk1/2, JNK1/2 and p38 kinases.

Figure 4.

 The MAPKs Erk1/2, JNK and p38 are not involved in P2X7R-induced cell death. Panel (a), analysis of phosphorylation of the Erk1/2, JNK1/2 and p38 MAPKs after P2X7R stimulation with 2.5 mM Bz-ATP, EGF or anisomycin during 15 min at 37°C. Proteins from NPC lysates were separated on 10% SDS/PAGE and immunoblotted for phospho-threonine and phospho-tyrosine with anti-Erk1/2 (pTpY 185/187) (first panel), anti-JNK1/2 (pTpY 183/185) (third panel), or anti-p38 (pTpY 180/182) (fourth panel) antibodies. After stripping, the three blots were probed with anti-total Erk1/2 (second panel) or anti-total p38 (last panel) antibodies to check for equal loading of gels. The results are typical of those obtained in three independent experiments. Panel (b), the effects of various inhibitors of MAPKs or MAPK-kinases on Bz-ATP-induced LDH release was investigated. NPCs were pre-incubated for 1 h at 37°C with 25 or 50 μM of U0126, a specific inhibitor of MEK1/2; SB203580, an inhibitor of the p38 MAPK; SP600125, an inhibitor of the JNK MAPKs. NPCs were then treated with 2.5 mM Bz-ATP for 3 h. The experiments were performed on separate days, and the values represent the mean and SD of three experiments.

To confirm that MAPK pathways are not involved in P2X7R mediated NPC death, we assessed the effect of specific inhibitors of Erk1/2 (U0126), JNK1/2 (SP600125) and p38 (SB203580) kinases on NPC death. No inhibition of LDH release was observed after pre-treatment of NPCs with any of the kinase inhibitors (Fig. 4b). Controls for the efficiency of the inhibitors were included (Fig. 4a). Thus, P2X7R induces NPC death independently of MAPKs activation.

P2X7R-activated non-selective pore formation is not observed in NPCs

Prolonged activation of P2X7R with ATP can lead to a huge increase in permeability revealing the formation of a non-selective pore, reportedly involved in cell death.

Following Bz-ATP treatment, the potential opening of a non-selective pore was assessed by measuring the uptake of EtBr in NPCs. P2X7R activation induced no dye uptake in NPCs (Fig. 5a) even after 60 min of treatment with Bz-ATP (data not shown). In contrast, significant EtBr uptake was observed in the mouse neuroblastoma cell line Neuro-2a (Fig. 5a). Similar results were observed in NPCs and Neuro-2a when YO-PRO-1 iodide was used instead of EtBr (data not shown). Thus, failure of NPCs to form the non-selective pore appears to be a feature of this cell population.

Figure 5.

 Neural progenitor cells (NPCs) do not form a non-selective pore after P2X7R stimulation. Panel (a), NPCs from wt and P2X7Rko mice and Neuro-2a cells were incubated with ethidium bromide (EtBr) and stimulated with 2.5 mM Bz-ATP for increasing lengths of time. EtBr fluorescence of each sample was monitored for 20 min at 37°C. For each curve, Bz-ATP-dependent EtBr uptake was calculated by subtracting EtBr values of unstimulated cells from values of Bz-ATP-stimulated cells. The results are representative of at least three experiments performed on different days. Panel (b), NPCs, astrocytes and macrophages from C57BL/6 mice were isolated and RNA was extracted as described in Materials and methods. Two micrograms of total RNA were reverse transcribed, and pannexin-1 expression was determined by RT-PCR.

Two hypotheses have been proposed to explain non-selective pore formation: (i) activation of a signalling pathway that leads to the opening of a pore formed by a protein distinct from P2X7R (North 2002; Faria et al. 2005) and (ii) a gradual permeability increase (dilation) of cationic channels (North 2002). The first hypothesis is supported by recent findings demonstrating that pannexin-1, a hemi-channel protein, mediates large pore formation following P2X7R activation in macrophages (Pelegrin and Surprenant 2006). By PCR, we found that NPCs as well as astrocytes and macrophages expressed pannexin-1 transcripts (Fig. 5b). Thus, the absence of P2X7R non-selective pore formation in NPCs is not due to a deficiency in pannexin-1 expression.

P2X7R-mediated cell death is calcium independent

P2X7 receptor is a Ca2+ permeable ionotropic receptor. As shown in Fig. 6(a), a rapid increase in intracellular Ca2+ is observed in NPCs triggered with Bz-ATP and not in NPCs from P2X7R-deficient mice. In addition, pre-treatment of NPCs with o-ATP inhibited Ca2+ influx after Bz-ATP stimulation (data not shown).

Figure 6.

 P2X7R-induced NPC death is calcium independent. Panel (a), NPCs from wt and P2X7Rko mice were loaded with 2 μM fura-2AM for 30 min and stimulated with 300 μM Bz-ATP in modified Krebs-HEPES at 37°C. The increase in Ca2+ was determined by dual excitation spectrofluorimetric analysis at 340 and 380 nm (ratio of OD340/OD380). Panel (b), NPCs from C57BL/6 mice were incubated with or without 50 μM BAPTA-AM, (Ca2+ chelator) for 1 h and loaded with 2 μM fura-2AM for the last 30 min and stimulated with 300 μM Bz-ATP in modified Krebs-HEPES with or without 5 mM EGTA (Ca2+ chelator) at 37°C. Panel (c), NPCs were incubated with 5 mM EGTA and 50 μM BAPTA-AM for 1 h and stimulated or not with 2.5 mM Bz-ATP for 3 h at 37°C. Percentage of LDH release was determined. The values represent the mean and SD of three different experiments performed on separate days.

As an increase in the intracellular Ca2+ concentration is involved in many types of cell death (Syntichaki and Tavernarakis 2003), we determined whether P2X7R-mediated cell lysis is calcium dependent in NPCs. In cells incubated with EGTA and/or BAPTA-AM, two Ca2+ chelator types, no rise in intracellular Ca2+ was found after P2X7R stimulation (Fig. 6b). Figure 6c shows that NPC lysis is induced by Bz-ATP in medium containing EGTA and/or BAPTA-AM. These results demonstrate that P2X7R-mediated NPC death is calcium independent.

NPC cell death is accompanied by loss of mitochondrial membrane potential

Mitochondria play a central role in the induction of death which is often the starting point for the activation of several signalling pathways that will each contribute to the elimination of the cell. Alterations of the mitochondrial membrane potential were visualized by flow cytometry, using the JC-1 dye. In Fig. 7(a), intact mitochondria (positive for JC-l aggregates) were detected in 57% of untreated NPC but in only 17% of NPC treated with Bz-ATP. In NPCs from P2X7R-deficient mice, Bz-ATP had little, if any, effect on the percentage of cells with a high mitochondrial membrane potential. It is worth noticing that the loss of mitochondrial potential is found 30 min after addition of Bz-ATP and correlates with plasma membrane disruption (Fig. 3c). To better characterize the kinetic of mitochondrial membrane depolarization of NPCs treated with Bz-ATP, we used the method described by Voronina et al. (2004). This method was used successfully by Garcia-Marcos et al. (2005) to study the mitochondrial membrane depolarization following P2X7R stimulation in cells from submandibular glands. As shown in Fig. 7(b), mitochondrial membrane depolarization of NPCs from C57BL/6 mice is rapidly induced after Bz-ATP addition and reaches a plateau after 25 min of incubation. In contrast, no depolarization is observed with NPCs from P2X7R ko animals. These data clearly establish that Bz-ATP stimulation of P2X7R triggers a rapid mitochondrial membrane depolarization in NPCs expressing this receptor.

Figure 7.

 P2X7R stimulation induces loss of mitochondrial membrane potential in NPCs. Panel (a), NPCs of wt and P2X7Rko mice were treated with 2.5 mM Bz-ATP in presence of 4.8 μM JC-1 dye during 30 min and analysed by flow cytometry. The results shown are representative of at least two experiments performed on different days. Panel (b), NPCs from wt and P2X7Rko mice were loaded with 1 μM TMRE for 30 min and stimulated with 1 mM Bz-ATP in modified Krebs-HEPES at 37°C. TMRE fluorescence of each sample was monitored for 30 min at 37°C. The results shown are representative of at least three experiments performed on different days.

Discussion

In this work, we show by RT-PCR and western blot that P2X7R is expressed in NPCs. Furthermore, as observed in several cell types, we find that P2X7R expressed at the plasma membrane of NPCs is present in similar amounts in DRMs and non-DRMs (Garcia-Marcos et al. 2006; Barth et al. 2007; Gonnord et al. 2009). Patch-clamp recordings showed that stimulation of P2X7R by Bz-ATP leads to the opening of this ionotropic purinergic receptor in NPCs. Prolonged activation of P2X7R with extracellular ATP or Bz-ATP leads to lysis/necrosis of NPCs accompanied by a loss of mitochondrial membrane potential. Surprisingly, NPC death mediated by P2X7R stimulation is not associated with the opening of a non-selective pore, which is assumed to be responsible for ionic imbalances leading to cell death.

The following points strongly support the conclusion that the observed biological effects are mediated by P2X7R activation: (i) high concentrations of ATP are needed to induce cell death, (ii) Bz-ATP induces NPC cell death with a greater potency than ATP (iii) the pharmacological inhibitor o-ATP which preferentially acts on P2X7R abolishes ATP-induced lytic death of NPC and (iv) ATP and Bz-ATP do not trigger cell lysis of NPC from P2X7R-deficient mice.

Activation of P2X7R can lead to cell death by apoptosis and/or lysis/necrosis, depending on the cell type (Zheng et al. 1991; Ferrari et al. 1999). P2X7R-mediated apoptosis may be linked to a signalling pathway implicating death receptors (DR). Denlinger et al. (2001) have reported that residues 436–531 of the C-terminal region of P2X7R exhibit sequence homologies with the death domain region of TNFR1, suggesting that P2X7R could interact with adaptator or effector proteins involved in caspase activation, a salient feature of apoptotic cell death. However, NPC express functionally inactive DR, whose binding by specific ligands does not induce apoptosis, because of the absence of caspase 8. In addition, in cytokine-treated NPCs, caspase 8 recruitment and activation are prevented by the high level of expression of death effector domain-containing protein PED/PEA15, which competes with caspase 8 for Fas-associated protein with death domain binding (Ricci-Vitiani et al. 2004). Moreover, NPCs express a high level of inhibitor of apoptosis protein, which inhibits tumor necrosis-related apoptosis-induced ligand-induced apoptosis in these cells (Peng et al. 2005). Activation of Fas receptors on NPCs induces proliferation rather than apoptosis (Ceccatelli et al. 2004). These findings suggest that proteins involved in DR-mediated cell death are not functional in NPCs and may explain why P2X7R mediates lysis/necrosis rather than apoptosis in NPCs.

Actually, necrosis was described more on morphological than biochemical criteria because contrary to apoptosis, its biochemical pathways are not as well defined. In thymocytes, P2X7R-induced necrosis/lysis has been shown to require the sequential activation of Src family tyrosine kinase(s), PI3 kinase, Erk1/2 and the proteasome (Auger et al. 2005). In NPCs, we found no evidence for the implication of this pathway, as none of the inhibitors of enzymes from this pathway affected NPC death (Fig. 5 and data not shown).

Mitochondrial membrane depolarization triggered by P2X7R stimulation in cells from submandibular glands was described by Garcia-Marcos et al. (2005). In agreement with their observations, we found that P2X7R activation induces a rapid mitochondrial membrane depolarization in NPCs (Fig. 7b). Mitochondria seem to play a central role in the induction of necrosis and can activate multiple death effectors. We have tested several inhibitors of biochemical pathways potentially triggered by mitochondrial dysfunction: antioxidant (butylated hydroxyanisol), protease inhibitors (Pepstatin, Tosyl phenylalanine chloromethyl ketone, N-Acetyl-Leu-Leu-Nle-CHO), inhibitor of necroptosis (Necrostatin), inhibitors of cytosolic phospholipase A2 [Arachidonyl trifluoromethyl ketone (AACOCF3), E-6-(bromomethylene)tetrahydro-3-(1-naphthalenyl)-2H-pyran-2-one haloenol lactone suicide substrate] (Syntichaki and Tavernarakis 2003). None of these inhibitors affects Bz-ATP induced NPC cell death (data not shown). These results suggest that several redundant pathways are activated simultaneously in the death process (Syntichaki and Tavernarakis 2003).

In thymocytes, P2X7R-mediated opening of the non-selective pore is required for cell death (Auger et al. 2005). However, in NPCs, P2X7R-stimulation results in lysis/necrosis of the cells without opening of the non-selective pore. Similarly, P2X7R stimulation does not seem to trigger pore formation in cells of neuronal origin. Indeed, in rat inner retina, ATP induces permeabilization only in microglia, which are cells from haematopoietic origin but not in rat retinal ganglion cells (Innocenti et al. 2004). In the latter work, the neuronal fate was not studied and it thus remains to be determined whether some of the cells undergo cell death, as NPCs. One may argue that our observations stem from the use of NPCs from C57BL/6 mice in which the P2X7R bears a Pro-451 to Leu mutation associated with a decrease in pore formation and cell death in mature T lymphocytes and thymocytes (Adriouch et al. 2002; Auger et al. 2005). Importantly, we found no evidence for a Bz-ATP-mediated dye-uptake pore in BALB/c mouse NPCs (data not shown) which indicates that the absence of pore formation is not a peculiarity of C57BL/6 NPCs.

Pannexin-1 has recently been shown to be involved in non-selective pore formation and IL-1β release triggered by P2X7R (Pelegrin and Surprenant 2006). This was established by silencing pannexin-1 with small-interfering RNA and pannexin-1 inhibitory peptide. We found that pannexin-1 is expressed in NPCs (Fig. 5b), therefore the absence of non-selective pore formation in NPCs does not appear to be due to a deficiency in pannexin-1 expression. Different studies suggest that the ability to form pores correlates with the density of cell surface P2X7R (Hickman et al. 1994; Narcisse et al. 2005). P2X7R expression increases during the differentiation of monocytes into macrophages (Hickman et al. 1994), as does the ability to form pores. Similarly, primary foetal human astrocytes treated with IL-1β showed increased surface P2X7R expression, which was associated with an increase in pore formation (Narcisse et al. 2005). We could hypothesize that immature neural cells do not express sufficient P2X7R sub-units to recruit proteins (such as pannexin-1) involved in non-selective pore formation or to generate critical amounts of a second messenger needed to open the pore (Faria et al. 2005). Thus, NPCs die by an original pathway that has not been described yet as it is independent of the non-selective pore and involves an undefined biochemical pathway.

During acute spinal cord injury in rats, high levels of ATP are released in the peritraumatic zone, and more cells die in areas of high ATP release than in regions with low ATP concentrations at the same distance from the lesion (Wang et al. 2004). In addition, it has been shown in EAE that pathogenic T lymphocytes of wild-type mice are eliminated in the CNS more efficiently than T cells of P2X7Rko animals. These experiments show that the amount of ATP needed to induce P2X7R-mediated cell death can be reached in the CNS during EAE (Chen and Brosnan 2006). Previous studies did not investigate NPC impairment because of ATP, but these cells are also likely to be targeted. Thus, endogenous ATP release from damaged cells could lead to cell death in vivo, thereby limiting endogenous repair and diminishing the efficiency of cell therapy by NPCs in neurodegenerative and demyelinating diseases as well as stroke.

Administration of P2X7R antagonists may represent a useful therapeutic approach to treat neurodegenerative disease, as it could inhibit cell death of repairing NPCs in vivo. In fact, o-ATP and pyridoxal-phosphate-6-azophenyl-2′,4′-disulfonate, antagonists of P2X7R, have been shown to improve recovery from spinal cord injury in rats (Wang et al. 2004) and o-ATP and Brilliant Blue G attenuate symptoms associated with chronic EAE in mouse (Matute et al. 2007). P2X7R antagonists could also diminish inflammation by inhibiting release of inflammatory cytokines, such as IL-1β, IL-18 and tumour necrosis factor-α, by microglia and macrophages. The recent development of stable P2X7R-selective antagonists may help in controlling inflammatory infiltrates and decrease the number of dying cells in the CNS.

Neural progenitor cells might represent a potent source of cells for cell-based therapy as they can be obtained from different tissues (embryonic, foetal and adult) and do not form teratocarcinoma, in vivo, unlike embryonic stem cells (Martino and Pluchino 2006). Transplantation of NPCs has also been shown to improve recovery from different CNS inflammatory diseases (Martino and Pluchino 2006). However, the therapeutic use of NPCs requires a profound understanding of the biological behaviour of NPCs. The present study indicates that delivery of NPCs to replace damaged cells or to improve repair could be restrained because of the high level of ATP released in the lesion.

Acknowledgements

This work was financed by Association de Recherche contre la Sclérose en Plaques (ARSEP), Fondation pour la Recherche Médicale and CNRS. CD was supported by ARSEP and Ligue Française contre la Sclérose en Plaques. We are very grateful to Drs C. Gabel (PGRD, Pfizer Inc.) for generously providing the P2X7R-deficient mice and G. Sadoc for providing the acquisition Elphy Software. We warmly thank Drs T. Amédée and G. Kroemer for their comments on the manuscript.

Ancillary