Mitochondrial Superoxide Generation Enhances P2X7R-mediated Loss of Cell Surface CD62L on Naïve Human CD4⁺ T Lymphocytes

Abstract

Migration of naïve CD4⁺ T lymphocytes into lymphoid tissue is essential for their activation and subsequent roles in adaptive immunity. The adhesion molecule CD62L, critical for this process, is highly expressed on naïve CD4⁺ T lymphocytes and is down-regulated upon T lymphocyte activation. We demonstrate protein expression of P2X7R on naïve CD4⁺ T lymphocytes and show functional channel activity in whole cell patch clamp recordings. CD62L down-regulation occurs rapidly in response to extracellular ATP, a process which is blocked by selective antagonists of P2X7R. This loss of surface CD62L expression was not associated with externalization of phosphatidyl serine. While investigating the mechanisms for this process we revealed that pharmacological modulation of mitochondrial complex I or III, but not inhibition of NADPH oxidase, enhanced P2X7R dependent CD62L down-regulation by increasing ATP potency. Enhanced superoxide generation in the mitochondria of Rotenone and Antimycin A treated cells was observed and may contribute to the enhanced sensitivity of P2X7R to ATP. P2X7R dependent exposure of phosphatidylserine was also revealed by pre-incubation with mitochondrial un-couplers prior to ATP treatment. This may present a novel mechanism whereby P2X7R dependent PS exposure occurs only when cells have enhanced mitochondrial ROS generation. The clearance of apoptotic cells may therefore be enhanced by this mechanism which requires functional P2X7R expression.

Introduction

The human ionotropic P2X7 receptor (P2X7R), cloned in 1997, has been shown using molecular and biochemical techniques to be highly expressed by immune cells of the hematopoietic linage, as well as erythrocytes (1-4). P2X7R expression has also been confirmed in a number of non-immune tissues including epithelial cells, which may up-regulate the receptor during inflammation, HUVECs and pancreatic duct cells (5-7). Activation of the receptor requires millimolar concentrations of its ligand ATP, which leads initially to the influx of Ca²⁺ and Na⁺ and the efflux of K⁺. This is followed by a second permeation state where activation and opening of Pannexin-1 hemichannels causes the uptake of molecules <900 Da and the activation of cellular death pathways (8,9).

The role of P2X7R in T lymphocyte biology has been investigated primarily in mice, where thymocytes and differentiated CD4⁺, CD8⁺ and CD4⁺CD25⁺ (Treg) T lymphocyte subsets were shown to undergo cell death following P2X7R activation (10-12). Studies of mice and human T lymphocytes have revealed that P2X7R may also be important for the activation of naïve T lymphocytes, their differentiation into effector lineages and subsequent proliferation (13-15).

Naïve T lymphocytes survey secondary lymphoid organs (SLO) as they circulate the body through the blood and lymphatic systems (16). If during this surveillance, naïve T lymphocytes come into direct contact with antigen presenting cells expressing both cognate antigen bound to major histocompatibility complexes and co-stimulatory molecules, then these cells undergo activation and clonal expansion. The homing to SLO and entry of naïve T lymphocytes by transendothelial migration through high endothelial venules, requires the expression of chemokine receptors such as CCR7 and adhesion molecules including CD62L (L-selectin) and integrins (17). Expression of these components is enriched on naïve and central memory T lymphocytes and is essential for the entry of these cells into SLO for their subsequent activation (18). Upon activation of these cells, CD62L is down-regulated and the pattern of chemokine receptors altered to allow egress from lymphoid organs and migration to peripheral sites of infection where the effector functions of these cells are required (16).

Studies initially with B lymphocytes isolated from human patients with B cell chronic lymphocytic leukaemia showed that ATP can induce CD62L down-regulation from the cell surface (19). This process was subsequently extended to T lymphocytes through studies of mouse cells and peripheral blood mononuclear cells isolated from healthy humans (20,21). The signalling mechanisms involved in CD62L down-regulation have been investigated in the context of CD3 and mitogenic activation. ADAM17 is the principle sheddase responsible for CD62L cleavage at the cell surface, although there are overlapping roles for ADAM10 in calcium induced CD62L loss (22,23). PKC and MAPKs directly phosphorylate ADAM17 and facilitate its trafficking to the surface and/or activation (24-26). The PI3K/mTOR pathway, through its Akt dependent translocation of FOXO1 from the nucleus, regulates the expression of KLF2 transcription factor-dependent genes including CD62L as well as CCR7 and S1P1 (27,28).

The mechanisms coupling P2X7R activation to CD62L down-relation are unclear and the aim of this study was to investigate established and novel signalling mechanisms in the regulation of CD62L. Previous studies have used RNA and protein techniques as well as well as dye uptake based assays, exploiting the second permeation state following P2X7R activation, to confirm P2X7R expression. Here we show P2X7R functional expression in naïve human CD4⁺ T lymphocytes using whole cell patch clamp electrophysiology as well as P2X7-dependent loss of surface CD62L expression. Remarkably, inhibitors of mitochondrial electron transport significantly enhanced potency of ATP/P2X7R-mediated CD62L down-regulation. PS exposure was not observed in response to ATP alone, however pre-treatment with Rotenone and Antimycin A revealed P2X7R dependent PS externalization. This suggests that modulation of mitochondrial function, which occurs during apoptosis, may promote clearance of apoptotic naïve CD4⁺ T lymphocytes in a P2X7R dependent manner.

Materials and Methods

Reagents

Unless otherwise indicated all reagents were purchased from Sigma Aldrich. GM6001 was purchased from Calbiochem (Merck), Darmstadt, Germany. A438079 and AZ11645373 were purchased from Tocris Bioscience, Bristol, UK. P2X7R and Erk1/2 antibody were purchased from Santa Cruz Biotechnology, Santa Cruz, California, USA. MitosoxRed and H₂DCFDA were purchased from Invitrogen, Paisley, UK. Annexin V/PI apoptosis kit was purchased from Southern Biotech, Atlanta, GA, USA.

Cells

Naïve CD4⁺ T lymphocytes were isolated from the peripheral blood of healthy volunteer donors using a naïve CD4⁺ T cell isolation kit II human (Miltenyi Biotec MACS). Procedures using human blood were carried out under University of Bath and Departmental safety and ethical guidelines for the use of human tissue. Freshly isolated naïve CD4⁺ T lymphocytes were cultured in RPMI 1640 (supplemented with 10% FCS, 10 μg/ml penicillin and 10 μg/ml streptomycin) and incubated in a 37°C 5% CO₂ incubator.

HEK293 cells were cultured in DMEM: F12 (supplemented with 10% FCS, 10μg/ml penicillin and 10μg/ml streptomycin) and incubated in a 37°C 5% CO₂ incubator. A P2X7R expression plasmid (Professor Alan North, The University of Manchester, UK) was transfected into confluent HEK293 cells using Lipofectamine 2000 reagent (Invitrogen, Paisley, UK) following manufacturers guidelines.

Immunoblotting

The cell stimulations, cell lysis, and Western blotting were performed as described previously (29). Cells were treated as stated, centrifuged 300g for 30 seconds, and lysed by addition of 100 μl solubilization buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 1% Nonidet P-40, 5 mM EDTA, 1 mM sodium vanadate, sodium molybdate, 10 mM sodium fluoride, 40 g/ml PMSF, 0.7 g/ml pepstatin A, 10 g/ml aprotinin, 10 g/ml leupeptin, 10 g/ml soybean trypsin inhibitor). The samples were mixed and gently rotated at 4°C for 20 min and then centrifuged at 13000g for 10 min. The supernatant was transferred to fresh tubes and diluted 4:1 with 10% SDS containing 5 × sample buffer. Before loading onto the gel, samples were boiled for 5 min at 100°C. The samples were separated by electrophoresis in 10% SDS-PAGE. Proteins were then electrotransferred onto nitrocellulose membrane, blocked in 5% milk, and incubated with anti–P2X7R (H-265) (1/1000 dilution; Santa Cruz #SC-25698) as primary Ab and anti-rabbit HRP (1/10,000 dilution) as secondary Ab. Immune complexes were visualized using ECL (ECL Western blotting system; Amersham Bio-science, Little Chalfont, U.K.).

Electrophysiology

Ionic currents were measured using the whole-cell patch clamp technique as previously described (30). Borosilicate glass GC150 TF-10 capillaries (Harvard Apparatus, Kent, UK) were pulled in two stages using a Narishige PC-10 and then fire-polished with a Narishige MF830 Microforge to give a filled resistance of between 3-5 MΩ. Internal recording solution contained: KCl (147 mM), Hepes (10 mM), EGTA (1 mM) and pH adjusted to 7.4 with KOH. External solution contained: NaCl (147 mM), KCl (2 mM), CaCl₂ (2 mM), MgCl₂ (1 mM), Hepes (10 mM), D-Glucose (12 mM) and pH adjusted to 7.4 with NaOH. Suspension cells were allowed to settle in the recording bath for at least 20 minutes. Recordings were performed under voltage clamp at the holding potential of -60 mV with a HEKA EPC10 amplifier using Patchmaster V2.11 software. All agonists and antagonists were applied using an RSC-200 rapid solution changer (BioLogic, France) built from GC100 T-10 glass capillaries (Harvard Apparatus, Kent, UK). Antagonists were applied 1 second prior to the subsequent co-application with the agonist. The external divalent cations Ca²⁺ and Mg²⁺ are known to inhibit P2X7R function therefore, for electrophysiology, unless stated otherwise agonists and antagonists were applied in external solution without MgCl₂ or CaCl₂ (31,32).

CD62L measurement by flow cytometry

For each experimental condition 0.5×10⁶ naïve CD4⁺ T lymphocytes were treated with vehicle/inhibitors before the addition of ATP at given concentrations. Cells were incubated in a 37°C 5% CO₂ incubator for the indicated times. Cells were washed twice with PBS + 2 % FCS then stained for 1 hour with either isotype control (IgGk1-FITC), CD62L-FITC or CCR7-APC on ice for 1 hour. After labeling, cells were washed a further two times and then analyzed by using a FACSCanto flow cytometer (BD, Oxford, UK). Cells were excited at the wavelength 488 nm and the emission wavelength recorded at 530/30 nm.

ROS generation detection

Freshly isolated naïve CD4⁺ T lymphocytes were incubated with 10 μM 2′,7′-dichlorodihydrofluorescein diacetate (H₂DCFDA, Invitrogen, Paisley, UK) in RPMI 1640 (without supplements) for 45 minutes at ambient room temperature protected from light. Cells were washed by centrifugation at 300g for 5 minutes. Cells were re-suspended in the same external solution as for electrophysiology 5 minutes prior to agonist application. Fluorescence was monitored using a multi-detection plate reader (Fluostar Optima, BMG Labtech, UK; excitation, 485 nm; emission, 520 nm). Measurements were performed in triplicate per treatment group. Linear regression was performed to determine the rate of ROS generation.

To measure mitochondrial O₂⁻ levels cells were loaded with 2.5 μM MitoSOX Red (Invitrogen, Paisley, UK) for 30 minutes at 37°C protected from light. Cells were then washed by centrifugation at 300g for 5 minutes, treated with agonist for the indicated time in complete RPMI 1640 cell culture medium, washed into PBS and analyzed immediately using flow cytometry. MitoSOX Red fluorescence was detected at the excitation and emission wavelengths of 488 nm and 585 nm respectively.

Measurement of phosphatidylserine (PS) externalization

Naïve CD4⁺ T lymphocytes were treated with extracellular ATP at the indicated concentrations in complete RPMI 1640. For each condition 0.25×10⁶ cells were washed twice in ice cold PBS and then re-suspended to a concentration of 1×10⁶ cells/ml in Annexin V binding buffer (10 mM HEPES pH 7.4, 140 mM NaCl, 2.5 mM CaCl₂, 0.1% Bovine Serum Albumin (BSA)). Using an Annexin V/PI apoptosis kit (Southern Biotech, Atlanta, GA, USA) allophycocyanin (APC) conjugated Annexin V was then added to each tube and cells incubated on ice protected from light for 15 minutes. Further Annexin V binding buffer was added and cells were immediately analyzed by flow cytometry at the excitation and emission wavelengths of 633 nm and 660 nm respectively.

Data analysis and statistics

For CD62L expression and MitoSOX Red, flow cytometry was analyzed using FACS Diva software (BD Biosciences, USA); a gate was set around naïve CD4⁺ T lymphocytes and mean fluorescence index (MFI) recorded for all cells in this gate. Alternatively, for measurement of PS exposure a second gate was set containing 1% of unstained cells positive for fluorescence; the percentage of fluorescence positive cells from treatment groups stained with antibody were recorded. Graphs were plotted and concentration/inhibition sigmoidal curves were fitted using Prism 4 (GraphPad, USA.); statistical analysis was performed using this software.

Results

P2X7R is expressed in T lymphocytes and functions as an ion channel

In order to explore further the function of P2X7R expression in human T lymphocytes, we first verified expression on primary human naïve CD4⁺ T lymphocytes freshly isolated from the blood. We also assessed the leukemic T cell line Jurkat as well as the monocyte leukemic cell line THP-1. Cell lysates were immuno-blotted with anti-P2X7R Ab to examine the expression of this receptor (Figure 1A). Specificity of the antibody for P2X7R was confirmed by comparing HEK293 cells transfected with vector only or a P2X7R plasmid.

Figure 1. P2X7R is expressed in T cell lines, primary activated and naïve CD4⁺ T lymphocytes.

A, Protein was isolated from 1×10⁶ primary human T cells or leukemic cell lines. SDS-PAGE was performed followed by immunoblotting with anti–P2X7R Ab. Naïve CD4⁺ cells donor shown is representative of samples from different donors collected on different days, leukemic cell lines are representative of 3 experiments. Antibody specificity was confirmed using HEK293 cells transfected with a P2X7R plasmid. Whole cell patch clamp electrophysiology: B, 5 mM ATP was applied for 10 seconds in buffer containing CaCl₂ and MgCl₂. C, 5 mM ATP was applied for 10 seconds in buffer nominally free from CaCl₂ and MgCl₂. D, Current density in presence and absence of divalent cations is compared. E, Changes in current density in response to increasing concentrations of ATP applied for 10 seconds, followed by 60 second pause, were measured. F, 5 mM ATP was applied for 10 seconds in the presence or absence of the P2X7R inhibitor A438079 (10 μM). G, Current density when ATP was applied in the presence or absence of 10 μM A438079 is compared. Bar represents the application of agonists and antagonists as indicated. Student paired t test performed for analysis of statistical significance, *p<0.05.

Having verified protein expression of P2X7R in naive CD4⁺ T lymphocytes freshly isolated from peripheral human blood we performed whole cell patch clamp recordings to evaluate functional ion channel expression. Previous studies have investigated the effect of divalent cations on P2X7R activation (3,31,32), concurrent with this we demonstrate that naïve CD4⁺ T lymphocytes exhibit a small inward current in response to extracellular application of 5 mM ATP (10 seconds) which is increased when MgCl₂ and CaCl₂ were removed from the application buffer (Figure 1 B-D). ATP is reported to act as an agonist for cloned human P2X7R stably expressed in HEK293 cells with an EC₅₀ value of 1.8 and ~0.7 mM, where external electrophysiological solution contained normal or low MgCl₂/CaCl₂ concentrations respectively (33,34), whereas the reported EC₅₀ for ATP interacting with other P2X receptors occurs with much lower values (35). Extracellular ATP evoked ionic currents in a concentration dependent manner with a response detected with 1 mM ATP suggesting the activation of P2X7R (Figure 1E). The sensitivity to a P2X7-selective competitive antagonist (A438079) was investigated to determine the contribution of P2X7R to ATP induced currents in T lymphocytes. A438079 inhibits activation of human P2X7R expressed in cell lines with an IC₅₀ value of 0.1-0.3 μM measured by calcium influx, large molecular weight dye uptake and IL-1β release (36). In this study, application of 10 μM A438079 significantly inhibited 5 mM ATP evoked currents in naïve CD4⁺ T lymphocytes (n=5 p<0.05) (Figure 1F-G) confirming the activation and functional expression of P2X7R.

P2X7R activation causes CD62L down-regulation

P2X7R has been shown to couple to CD62L down-regulation in both mouse and human T lymphocytes (20). Here we further explore the mechanism integrating ATP signaling through P2X7R to CD62L processing. Naïve CD4⁺ T lymphocytes isolated from peripheral human blood show uniform high levels of CD62L surface expression as well as displaying a naive expression pattern of other markers including CCR7 (Supplementary figure 1 A-C). PMA through activation of PKC is known to cause rapid potent CD62L down-regulation (37). Indeed, stimulation of freshly isolated human naïve CD4⁺ T lymphocytes with 100 nM PMA for 30 minutes induced significant CD62L down-regulation. Treatment with 3 mM ATP for 1 hour also caused significant CD62L down-regulation (Figure 2A and Supplementary figure 1 D). The level of ATP induced CD62L down-regulation after 1 hour 3 mM treatment varied significantly between donors whereas TCR induced loss was less variable (Supplementary figure 1E). ATP induced CD62L down-regulation was rapid with a peak loss after 15 minutes and sustained low surface expression for up to 6 hours (Figure 2B). We next investigated temporal relationship between ATP treatment and CD62L down-regulation (Figure 2C). Concentration response curves show EC₅₀ values of 877 μM (5 minute stimulation) and 888.6 μM (1 hour stimulation) suggesting the sustained involvement of P2X7R over time (38). To confirm P2X7R function in this process using available pharmacological tools we chose two P2X7R antagonists: A438079 (which blocks ATP induced currents in Figure 1) and a non-competitive P2X7R antagonist AZ11645373 (Figure 2D). Published data indicate that AZ11645373 is more potent than A438079 with IC₅₀ values in the range of 5-90 nM at human P2X7R expressed in HEK293 cells and in THP-1 monocytes (33). Pre-treatment for 30 minutes with both antagonists inhibited CD62L down-regulation in response to 1 hour 3 mM ATP treatment with IC₅₀ values of 2.25 μM (A438079) and 1.35μM (AZ11645373). In naïve CD4⁺ T lymphocytes these antagonists act with an order of magnitude less potency than figures published for HEK293 cells transfected with human P2X7R (33,36).

Figure 2. ATP induces CD62L down-regulation from the surface of naïve CD4⁺ T lymphocytes through P2X7R.

A, Naïve CD4⁺ T lymphocytes (1×10⁶ cells/ml) were treated for 30 mins with either vehicle or PMA (100 nM) or for 1 hour with ATP (3 mM). Cell surface CD62L expression was measured by flow cytometry. B, To measure the kinetics of loss of CD62L surface expression, cells were treated with ATP (3 mM) for the times indicated. C, Cells were treated with increasing concentrations of ATP as indicated for 5 minutes and 1 hour before measuring CD62L expression. D, Inhibition curves displaying the effect of 30 minute A438079 and AZ11645373 pre-treatment on 3 mM ATP-induced (1 hour) CD62L down-regulation. E, Cells were pre-treated for 30 minutes with a broad spectrum MMP antagonist GM6001 before addition of ATP (3 mM) for one hour. Data are the mean of at least 3 independent experiments using cells from different donors ± SEM. One Way ANOVA, followed by Tukey’s post-hoc test was performed to compare significance differences between treatment groups # p<0.05, ** p< 0.01, *** p<0.001.

We initially used small molecule kinase inhibitors to look for signaling molecules involved in ATP induced CD62L loss, however, we report that PI3K, MEK-Erk1/2 and PKC signaling are not required for this process (Supplementary figure 2). PI3K and MEK-Erk1/2 signaling have been shown to be dispensable for ATP induced processing of CD27 in mouse lymphocytes (39). We reasoned that, as P2X7R activation leads to calcium influx (40) and CD62L processing can be activated by raising intracellular free Ca²⁺ (22), ATP induced CD62L down-regulation might be dependent on influx of Ca²⁺ through P2X7R. While elevation of cytosolic Ca²⁺ by thapsigargin (1-100 μM) caused significant down-regulation of CD62L, the absence of presence of calcium had no significant impact on ATP-induced CD62L down-regulation (n=3 p=0.9736) (Supplementary figure 3). This is consistent with previous studies which have shown Ca²⁺ influx is not necessary for ATP induced CD62L down-regulation (19). These observations indicate that, whilst increases in cytosolic Ca²⁺ levels can cause CD62L down-regulation, ATP-induced loss of CD62L in naïve CD4⁺ T lymphocytes occurs via a calcium-independent mechanism.

ADAM17 is the principle proteinase responsible for CD62L cleavage in response to a number of activating factors, but recently evidence has suggested that P2X7R mediated CD62L down-regulation also occurs through ADAM10 activation (19). When naïve CD4⁺ T lymphocytes were pre-treated with GM6001 (100 μM), a broad spectrum Matrix Metallo-Proteinase (MMP) inhibitor, significant inhibition of CD62L down-regulation in response to 1 hour 3 mM ATP was observed (n=3 p<0.05) (Figure 2E).

ATP couples to reactive oxygen species generation

P2X7R activation can lead to generation of intracellular ROS and this drives biochemical processes within cells (41). However, little is known about the role of P2X7R in ROS generation in T lymphocytes. ADAM17 can be activated by ROS through oxidation of cysteine motifs (42) and we therefore hypothesized that ROS generation in response to ATP could activate ADAM17 and subsequently lead to CD62L processing. Treatment of naïve CD4⁺ T lymphocytes with ATP caused an increase in the rate of ROS generation compared to vehicle alone (Figure 3A). With 5 mM ATP treatment, a significant increase in the rate of DCF fluorescence was observed (n=3 p<0.05). DCF detects a variety of intracellular ROS species including: H₂O₂, hydroxyl radicals, peroxyl radicals, ONOO⁻ and NO. Previous studies have reported H₂O₂ to cause CD62L down-regulation (42) and in naïve CD4⁺ T lymphocytes we observed this effect to be rapid upon stimulation with 100 μM H₂O₂ (Figure 3B).

Figure 3. In ATP treated cells, ROS generation is increased and un-coupling of mitochondrial electron transport chain complex I and III causes enhanced CD62L down-regulation.

A, Naïve CD4⁺ T lymphocytes (1×10⁶ cells/ml) loaded with 10 μM DCF were treated with increasing concentrations of ATP for 1 hour and the rate of change of DCF fluorescence was monitored as described in Materials and Methods. B, Cells were treated with H₂O₂ (100 μM) for increasing periods of time and CD62L surface expression was measured. Naïve CD4⁺ T lymphocytes (1×10⁶ cells/ml) were pre-treated with vehicle (DMSO), DPI (C) for 1 hour or Apocynin (D), Rotenone (E) or Antimycin A (F) for 30 minutes. Cells were then treated with 3 mM ATP for 1 hour before cell surface CD62L expression was measured by flow cytometry. Data are the mean of at least 3 independent experiments using cells from different donors ± SEM. One Way ANOVA, followed by Tukey’s post-hoc test was performed to compare significance differences between treatment groups */# p<0.05, ***/### p<0.001.

We employed a number of small molecule inhibitors of ROS generating enzymes to determine if ROS generation through ATP caused CD62L down-regulation. Unexpectedly, pre-treatment with DPI, an inhibitor of flavone containing enzymes which include NADPH oxidase and complex I of the mitochondrial respiratory chain, caused a significant enhancement of ATP induced CD62L down-regulation (Figure 3C). Remarkably, the NADPH oxidase inhibitor Apocynin had no effect on ATP induced CD62L loss suggesting the enhancing DPI effect is independent of NADPH oxidase (n=3 p>0.05) (Figure 3D). Rotenone, an un-coupler of mitochondrial electron transport at complex I, also enhances ATP induced CD62L down-regulation in a concentration dependent manner, suggesting DPI is acting through complex I (Figure 3E). Interestingly, pre-treatment with Rottlerin, a non-specific PKC inhibitor which also modulates ROS generation (43) also caused significant increase in ATP induced CD62L down-regulation (Supplementary figure 2 D). Antimycin A inhibits complex III and pretreatment with this compound also led to a significant enhancement of ATP induced CD62L loss (Figure 3F). Importantly, DPI, Rotenone and Antimycin A had no significant effect on basal CD62L surface expression and the P2X7R inhibitor A438079 inhibited their enhancing effect on the ATP response (Figure 4 A and B). This suggests that these compounds require P2X7R activation to alter CD62L surface expression. ROS can affect the function of a number of ion channels including P2X2R and we postulated that Rotenone and Antimycin A may be affecting P2X7R sensitivity (44,45). To confirm this we treated naïve CD4⁺ T lymphocytes with increasing concentrations of ATP following pre-treatment with vehicle (DMSO), 5 μM Rotenone or 1 μM Antimycin A and measured CD62L expression (Figure 4C). We observed a leftward shift in the concentration response curves for ATP with Rotenone and Antimycin A compared to DMSO pre-treatment. This indicates an increase in the potency of ATP in the presence of Rotenone and Antimycin A, as reflected by the lower EC₅₀ for ATP in the presence of Rotenone and Antimycin A (both 0.43 mM) compared to DMSO (1.58 mM). This provides evidence that that these compounds enhance ATP potency in the process of CD62L down-regulation.

Figure 4. Rotenone and Antimycin A enhance ATP induced CD62L down-regulation and increase mitochondrial O₂⁻ production.

Naïve CD4⁺ T lymphocytes (1×10⁶ cells/ml) were pre-treated with DMSO or A438079 (10 μM) and A, Antimycin A (1 μM) or B, Rotenone (5 μM). Cells were then treated with ATP (3 mM) for 1 hour and CD62L surface expression measured using flow cytometry. C, Cells were pre-treated with DMSO, Antimycin A (1 μM) or Rotenone (5 μM) for 30 minutes prior to addition of ATP at the concentrations indicated for 1 hour, after which CD62L expression was measured using flow cytometry. Cells were loaded with 2.5 μM MitoSOX Red, as described in Materials and Methods, and superoxide production was measured after 90 minutes treatment with D, Rotenone or E, Antimycin A at the concentrations indicated. Data are the mean of at least 3 independent experiments using cells from different donors ± SEM. One or Two Way ANOVA, followed by post-hoc tests was performed to compare significance differences between treatment groups * p<0.05, **/## p< 0.01, ***/### p<0.001.

Uncoupling of complex I and III from mitochondrial electron transport chain causes enhanced O₂⁻ generation

Mitochondrial electron transport under normal physiological conditions causes the leakage of a small number of electrons which can react with O₂ to form the ROS O₂⁻. In diseases driven by mutations to mitochondrial DNA (mtDNA) or where increased ROS generation cause mitochondrial damage, this mitochondrial O₂⁻ generation can significantly increase. Uncoupling of mitochondrial electron transport at complex I and III has been shown to cause significantly enhanced O₂⁻ generation (46,47). This led us to investigate mitochondrial O₂⁻ generation as a possible mechanism of the enhancing effect of Rotenone and Antimycin A. O₂⁻ levels were measured using the dye MitoSOX Red which is targeted to the mitochondria and fluoresces when oxidized by O₂⁻. Treatment with Rotenone or Antimycin A caused a concentration dependent increase in mitochondrial O₂⁻ generation (Figure 4D and E p<0.001).

Rotenone and Antimycin A reveal P2X7 dependent PS externalization

P2X7R activation in mouse CD4⁺ T lymphocytes has been linked to externalization of PS which is normally confined to the inner leaflet of the plasma membrane. We did not observe significant PS “flopping” in response to 1 hour 3 mM ATP treatment, however, pre-treatment with 5 μM Rotenone or 1 μM Antimycin A followed by ATP treatment induced a significant P2X7R dependent increase in PS surface exposure (Figure 5). This indicates that the effect of Rotenone and Antimycin A on P2X7R function is not limited to CD62L down-regulation. Previous studies with murine T lymphocytes have observed shrinkage of cells following ATP treatment (48,49); however we did not observe this response in human naïve CD4⁺ T lymphocytes (Supplementary Figure 4 B). Additionally, treatment of cells with ATP did not lead to necrotic cell death (Supplementary Figure 4 C), measured by LDH release, and pre-treatment with Rotenone did not enhance LDH release from ATP treated cells (Supplementary Figure 4 D). These data suggest that mitochondrial perturbation enhances the apoptotic marker PS, but does not cause cells to undergo necrotic cell death.

Figure 5. Rotenone and Antimycin A also enhance phosphatidyl serine (PS) exposure in response to ATP.

Naïve CD4⁺ T lymphocytes (1×10⁶ cells/ml) were pre-treated with DMSO, Rotenone (5 μM) or Antimycin A (1 μM) for 30 minutes prior to addition of ATP (3 mM) for 1 hour. Cells were then analysed for PS exposure by Annexin V binding as described in Material and Methods. Data are the mean of at least 3 independent experiments using cells from different donors ± SEM. Two Way ANOVA followed by post-test to compare treatment groups ** p< 0.01.

Discussion

In this study, we demonstrate the expression of P2X7R in human naïve and activated CD4⁺ T lymphocytes at the protein level. This receptor is functional as demonstrated by concentration-dependent ATP-induced inward currents and loss of surface CD62L expression. The ATP responses occurred in response to concentrations in the low millmolar range that would be expected to activate P2X7R. Indeed, both ATP-stimulated inward current and loss of surface CD62L were sensitive to pre-treatment with P2X7R antagonists. Remarkably, the loss of CD62L surface expression was insensitive to inhibitors targeting PKC, PI3K, MEK/ERK-1/2, molecules known to be involved in CD62L shedding induced by other agents (25-28). A broad-spectrum MMP antagonist inhibited loss of CD62L surface expression, suggesting that these proteases might be involved in shedding of this receptor in response to ATP/P2X7R stimulation. Some MMPs such as ADAM17 can be activated by ROS, which we demonstrated is elevated in naïve CD4⁺ T cells following ATP treatment. Remarkably, use of small molecule inhibitors of mitochondrial electron transport chain such as Rotenone and Antimycin A led to a significant enhancement of mitochondrial O₂⁻ and oxidative stress which correlated with enhanced sensitivity to ATP. Both Rotenone and Antimycin A enhanced ATP-induced CD62L down regulation. Co-treatment with Rotenone or Antimycin A with ATP also led to P2X7-dependent PS externalization which did not occur with individual treatments.

P2X7R expressed on human naïve CD4⁺ T lymphocytes acts as a functional ion channel, sensitive to divalent cations and inhibition by A438079. The inward current is initially rapid followed by a second slower phase, when ATP is removed the current returns to resting levels after a short delay. This current profile is similar to that observed in Xenopus oocytes expressing P2X7R cloned from human B lymphocytes (50). Previous studies using voltage clamp electrophysiology investigated the divalent cation sensitivity of rat and human P2X7R expressed in HEK293 cells (3,32). When cells were treated with 30 μM BzATP, Ca²⁺ and Mg²⁺ inhibited 50% of currents at 3.2 and 2.2 mM respectively. A number of processes downstream of P2X7R activation by ATP are blocked by the presence of extracellular Mg²⁺ ions, including cell shrinkage and down-regulation of cell surface CD62L and CD23 expression (19,49,51). Whilst removal of CaCl₂ and MgCl₂ from the extracellular solution potentiated P2X7R mediated currents in this study, the absence of these divalent cations from the extracellular solution did not affect CD62L down-regulation.

The antibody used to detect P2X7R expression in this study recognizes amino acid residues 331-595 of the C-terminus of P2X7R. A recent study revealed that human P2X7R, like P2X4R, is alternatively spliced and three of these eight isoforms have a truncated C-terminal domain (52,53). Consequently, the anti-P2X7R antibody used here can potentially only recognize five of these isoforms including the originally identified form, sometimes referred to as P2X7A receptor. PCR revealed that P2X7A is expressed in resting human CD4⁺ T lymphocytes along with P2X7B, which has a truncated C-terminus (52). P2X7B has been cloned and its function when expressed in HEK293 cells has been explored when expressed either alone or in combination with P2X7A (52,53). While P2X7B alone does not couple to pore formation or caspase activation, it can form heterotrimeric structures with P2X7A which causes enhanced pore formation. In naïve CD4⁺ T lymphocytes we observe significant ethidium bromide incorporation (a measure of pore formation) in response to ATP which is insensitive to inhibition by A438079 (Supplementary figure 4A). This may indicate that the pore formation in human naïve CD4⁺ lymphocytes responding to ATP involves either splice variants of P2X7R and/or other purinergic receptors that are insensitive to A438079.

Several in vitro studies have investigated the potential source of endogenous extracellular ATP that may contribute to lymphocyte and immune cell activation in vivo (54,55). Recent evidence suggests that following T lymphocyte activation, ATP is released through Pannexin-1 channels and acts in an autocrine manner to activate P2X receptors (13,14). Indeed, a recent study showed that the bee venom component melittin causes release of ATP leading to P2X7R dependent cleavage of E-cadherin and EGFR ligand from keratinocytes via ADAM10 and 17 respectively (56). Additionally, cardiac fibroblasts subjected to hypotonic stress, release ATP through connexin channels, which in turn, acts on P2Y2 receptors in an autocrine manner to cause the release of pro-fibrotic factors (57). The study of leukemic Jurkat T cells undergoing apoptosis revealed a novel mechanism of Pannexin-1 activation that involves C-terminal cleavage of this hemi-channel by caspases (58). This could represent a mechanism where ATP acts as a “find-me” signal to promote the clearance of apoptotic cells, through recruitment of phagocytes (59). ATP could potentially reach millimolar concentrations locally when released from Pannexin-1 channels in an autocrine or pancrine manner from healthy or apoptotic cells respectively. In vitro techniques to measure concentrations of ATP release following stimulation have been developed (54) and techniques are now emerging to measure in vivo levels of external ATP (60). It will be critical to measure external ATP concentrations in SLOs and the periphery under normal and inflammatory conditions to further understand the contribution of lymphocyte P2X7R to inflammatory responses.

Pharmacological un-coupling of mitochondrial electron transport at complex I and III leads to oxidative stress that enhances lymphocyte P2X7R downstream signalling events. This is consistent with previous reports that inhibition of complex I or III by Rotenone and Antimycin A/Myxothiazol could enhance P2X2R activation and which have implicated ROS in ion channel modulation (44,45). The impact on P2X7R responsiveness to ATP could occur at two levels. First, oxidative stress may lead to direct effects on P2X7 similar to that observed in P2X2R where ROS potentiate receptor activity through intracellular C terminal Cys430 residue (44). Indeed, P2X7R has a cysteine rich C terminal domain that is potentially sensitive to modulation by oxidative stress that leads to the potentiation of receptor activation (61). Alternatively the site of interaction could be down-stream of P2X7R activation and involve intermediary proteins regulated by P2X7R activation and sensitive to ROS. One such protein could be the stress-activated MAPK p38, which is phosphorylated in response to ATP (62). Indeed, in human monocytes, LPS and H₂O₂ induce ADAM17 activation through the p38 MAPK signalling pathway (25,63).

Un-coupling of mitochondrial electron transport at complex I and III in the presence of millimolar ATP also leads to significant PS exposure. In this study the role of P2X7R in these processes was confirmed using the next generation antagonists A438079 and AZ11645373, which show improved selectivity compared to previous P2X7R inhibitors (64). PS is normally confined to the inner leaflet of the plasma membrane, cells undergoing apoptosis externalise PS which acts as cue for phagocytes to engulf and destroy apoptotic material (65,66). Interestingly, neither ATP alone nor un-coupling of complex I or III in the absence of ATP led to PS externalization. This is contrast to studies of murine CD4⁺ T lymphocytes which show significant PS externalization in response to ATP alone (67,68). The ability of mitochondrial O₂⁻ to modulate P2X7R function may represent a novel protective mechanism. Cells which are under physiological conditions lose CD62L expression, but do not externalize PS in response to extracellular ATP and can presumably function as normal. We hypothesize that during T lymphocyte activation, ATP is released via Pannexin-1 channels to act in an autocrine manner to facilitate CD62L down-regulation and allow egress of cells from SLOs. In the inflamed periphery, ATP released from damaged cells could lead to down-regulation of CD62L from naïve or memory CD4⁺ T lymphocytes and prevent their entry into the lymphatic system. In contrast cells undergoing oxidative stress, that are also exposed to high levels of extracellular ATP, respond by externalization of PS which may act as a “find-me signal” for efficient removal from the body by phagocytosis. We propose a model where P2X7R activation, under conditions of oxidative stress, adds to the resolution of inflammation under physiological or pathological conditions. Indeed P2X7R has been implicated in a number of inflammatory and autoimmune disorders including: rheumatoid arthritis, multiple sclerosis, ALS and SLE (69-73). Although these studies have been primarily in mouse models, where P2X7R may not be sensitive to modulation by ROS, evidence exists to suggest that P2X7R may also play a role in these diseases in humans (74). Interestingly, the pathobiology of some of these diseases involves a mitochondrial ROS component.

However the contribution of lymphocyte P2X7R to inflammatory responses is complicated by the expression of SNPs that either increase or inactivate P2X7R function (75). The expression of P2X7R SNPs varies between individuals leading to differences in P2X7R function (76), indeed, expression of SNPs can determine the clearance of M. tuberculosis by macrophages. Additionally, the relative endogenous co-expression of full length P2X7A compared to inactivating and enhancing P2X7R splice variants will influence ATP mediated processes (52,53). Indeed, we have observed that the magnitude of CD62L down-regulation following ATP treatment can vary between healthy human donors (Supplementary figure 1E); this variation also extends to other processes such as ROS generation (Data not shown). Inactive P2X7R would be unlikely to elicit this putative protective mechanism and as a consequence, cells damaged by excessive oxidative stress would avoid clearance and potentially enhance inflammation. It will be important to evaluate the contribution of P2X7R in conditions of oxidative stress to the overall inflammatory response particularly in inflammatory and autoimmune conditions associated with changes in mitochondrial function.

In summary, we present a novel mechanism involving modulation of P2X7R activity by mitochondrial oxidative stress that may be involved in the resolution of inflammation (Figure 6). This protective mechanism should be considered in diseases where CD4⁺ T lymphocytes are exposed to increased ROS. In addition, the relative expression of P2X7R SNPs and splice variants in lymphocytes of inflamed patients may be of relevance in this model.

Figure 6. Effect of uncouplers of complex I and III of the mitochondrial electron transport on ATP induced loss of cell surface CD62L expression and PS exposure.

1. ATP induces loss of cell surface CD62L expression through P2X7R. 2. Rotenone and Antimycin A are uncouplers of complex I and complex III of the mitochondrial respectively. 3. Rotenone and Antimycin A significantly increase mitochondrial superoxide levels. 4. Pre-treatment with these compounds or DPI, an inhibitor of flavin containing enzymes such as complex I, but not Apocynin (5), an NADPH oxidase inhibitor, significantly enhance ATP induced CD62L down-regulation. 6. Pre-treatment with Rotenone or Antimycin A reveals P2X7R dependent externalisation of PS.