The P2X7 receptor (P2X7R) is an ATP-gated cation channel that activates caspase-1 leading to the maturation and secretion of IL-1β. Because previous studies indicated that extracellular Cl− exerts a negative allosteric effect on ATP-gating of P2X7R channels, we tested whether Cl− attenuates the P2X7R→caspase-1→IL-1β signaling cascade in murine and human macrophages. In Bac1 murine macrophages, substitution of extracellular Cl− with gluconate produced a 10-fold increase in the rate and extent of ATP-induced IL-1β processing and secretion, while reducing the EC50 for ATP by 5-fold. Replacement of Cl− with gluconate also increased the potency of ATP as an inducer of mature IL-1β secretion in primary mouse bone marrow-derived macrophages and in THP-1 human monocytes/macrophages. Our observations were consistent with actions of Cl− at three levels: 1) a negative allosteric effect of Cl−, which limits the ability of ATP to gate the P2X7R-mediated cation fluxes that trigger caspase-1 activation; 2) an intracellular accumulation of Cl− via nonselective pores induced by P2X7R with consequential repression of caspase-1-mediated processing of IL-1β; and 3) a facilitative effect of Cl− substitution on the cytolytic release of unprocessed pro-IL-1β that occurs with sustained activation of P2X7R. This cytolysis was repressed by the cytoprotectant glycine, permitting dissociation of P2X7R-regulated secretion of mature IL-1β from the lytic release of pro-IL-1β. These results suggest that under physiological conditions P2X7R are maintained in a conformationally restrained state that limits channel gating and coupling of the receptor to signaling pathways that regulate caspase-1.
Extracellular nucleotide homeostasis and P2 nucleotide receptors play important roles in regulation of the innate immune system (1, 2, 3). Particular attention has focused on the ability of the P2X7 receptor (P2X7R) to trigger proteolytic maturation and secretion of the proinflammatory cytokines IL-1β and IL-18 from macrophages. Several unique characteristics distinguish the P2X7R from other ionotropic P2X or metabotropic P2Y nucleotide receptors. Although most P2X receptors are largely expressed in excitable tissues and are involved in neural signaling, P2X7R is predominantly expressed in hemopoietic tissues (3). Activation of P2X7R requires millimolar concentrations of agonist (EC50 for ATP ≈500 μM), which may be found at “immune synapses” or in the context of cell lysis or bacterial invasion. In response to activation, P2X7R facilitates an immediate flux of monovalent and divalent cations, whereas prolonged stimulation of the receptor induces the flux of molecules ≤800 Da through a nonspecific pore (4, 5). P2X7R stimulation additionally activates the small GTPase Rho, which in turn, can activate phospholipase D, Rho-effector kinase (ROCK),3 and ROCK-dependent membrane blebbing (6, 7, 8). Depending on the magnitude and duration of stimulation, P2X7R can also elicit necrotic lysis or apoptotic death of activated leukocytes (3).
P2X7R-dependent IL-1β secretion remains incompletely understood. IL-1β is expressed as a 31–33 kDa inactive procytokine in the cytosol in response to inflammatory stimuli (such as LPS) that activate NF-κB target gene transcription (9). Activation and secretion of pro-IL-1β depends on cleavage to its 17-kDa mature IL-1β (mIL-1β) form by caspase-1, which itself exists as a low activity zymogen (procaspase-1). P2X7R activation induces a complete collapse of cation gradients that switches the cytosol from a high K+/low Na+ ionic milieu to a low K+/high Na+ environment. Previous studies have demonstrated that these ionic fluxes induce the rapid cleavage of procaspase-1 to its mature tetrameric (p10/p20) form (10, 11). This altered ionic milieu presumably triggers association of procaspase-1 with other components of the IL-1β processing machinery (such as apoptosis-associated Speck-like protein containing a CARD (ASC) and NACHT and leucine-rich region protein NALP1 and NALP3) as part of the recently identified inflammasome complexes (12, 13). Caspase-1 and mIL-1β are then rapidly secreted via an atypical and poorly defined mechanism (9). P2X7R knockout mice are resistant to the development of collagen-induced arthritis, have excess osteoclast activity, exhibit a greatly reduced sensitivity to chronic inflammatory pain, and express macrophages that do not release mIL-1β in response to ATP; these findings have unequivocally demonstrated the role of the P2X7R in innate immunity and inflammatory biology (14, 15, 16, 17).
The ability of the P2X7R to facilitate this rapid and massive export of IL-1β has been analyzed at physiological and pharmacological levels (18, 19, 20, 21, 22, 23). Macrophages from ASC knockout mice can neither process nor release mIL-1β (or IL-18) in response to P2X7R activation (23). P2X7R-induced secretion of mIL-1β is blocked by a range of chemically diverse pharmacological agents including glyburide, which inhibits ATP binding cassette transporter types (18, 19); AG126, a tyrphostin family tyrosine kinase inhibitor (20); diaryl sulfonylureas, which interact with GST-omega 1-1 (21); and bromoenol lactone, which blocks calcium-independent phospholipase A2 (10). The exact components of the caspase-1/IL-1β processing machinery targeted by these inhibitors are not known. In an analysis of P2X7R-dependent plasma membrane dynamics for insights into IL-1β release mechanisms, we demonstrated that P2X7R activation leads to rapid and dramatic membrane blebbing dependent on activation of Rho and ROCKs (6). However, this blebbing could be repressed by ROCK inhibitors without disruption of P2X7R-dependent IL-1β secretion. Mackenzie et al. (24) have provided microscopic and biochemical evidence for a mechanism of IL-1β compartmentation within evaginating microvesicles (≤0.5 μM in diameter) that are shed in response to P2X7R activation. In contrast, Andrei et al. (25) reported that ATP stimulates caspase-1 processing of pro-IL-1β to mIL-1β within a subset of secretory lysosomes, which release their contents via exocytosis due to alterations in ionic homeostasis and activation of phospholipases.
Earlier research suggested that most agents, such as ATP, that induce release of IL-1β additionally trigger cytolysis of activated monocytes/macrophages (26). This suggestion has raised questions as to whether cytolysis is a necessary component of the signaling cascade by which P2X7R activates the externalization of biologically active IL-1β. Ferrari et al. (27) reported that P2X7R-elicted IL-1β secretion was not necessarily associated with cell death because concentrations of ATP sufficient for IL-1β secretion were unable to initiate significant lactate dehydrogenase (LDH) release. However, Le Feuvre et al. (28) used macrophages from caspase-1 or P2X7R knockout mice to show that LPS priming of macrophages to induce pro-IL-1β expression also rendered macrophages more sensitive to the cytolytic effects of ATP, such that shorter ATP pulses were able to induce significant, caspase-1-dependent LDH release. Another recent report indicated that neutrophil-derived antimicrobial peptides (protegrins) are capable of inducing a loss of membrane integrity in monocytes, coupled with release of both pro-IL-1β and mIL-1β (29). Furthermore, cytolysis per se is recognized as a regulated cellular process, with a defined sequence of processes, which can be pharmacologically manipulated. This process has been well-studied using the dinoflagellate toxin maitotoxin, which induces the formation of a pore similar to that induced by the P2X7R and whose cytolytic actions are inhibited by the cytoprotectant glycine (4, 30). We recently demonstrated that maitotoxin treatment of LPS-primed macrophages leads to both regulated processing and secretion as well as cytolytic release of IL-1β (30).
How the P2X7R→caspase-1→IL-1β signaling cascade is triggered at in vivo inflammatory loci is not understood. In particular, the source and amount of ATP within such extracellular compartments has not been directly measured. Most nucleotide receptors, including the other six subtypes of the P2X receptor gene family, can be maximally activated by nanomolar to micromolar concentrations of extracellular ATP (2, 31, 32). In contrast, P2X7R is characterized by an exceptionally low affinity for ATP with EC50 values in the millimolar range (31). Thus, as typically analyzed under in vitro tissue culture conditions, significant activation of P2X7R and caspase-1 requires stimulation of isolated macrophages for at least 5 min with high (>1 mM) concentrations of extracellular ATP. It remains unknown whether such high extracellular ATP levels can be achieved and sustained within the interstitial spaces of inflamed tissues except perhaps during massive lysis of host cells or rapid killing of invading pathogens. The unusually low affinity of P2X7R for ATP has prompted speculation that the ATP affinity and/or conformational state of the receptor may be potentiated by in vivo conditions that are not effectively replicated in tissue culture environments.
Previous electrophysiological and flow cytometry studies indicated that the ability of ATP to gate P2X7R channel opening and pore formation can be strongly modulated by the composition of the extracellular medium (33, 34). In particular, substitution of Na+ with other cations, or replacement of Cl− with organic anions, increased the sensitivity of the P2X7R to ATP by an order of magnitude. In this study, we have tested whether manipulation of extracellular Cl− concentration similarly affects the ability of P2X7R to activate caspase-1, IL-1β processing, and secretion and cytolysis in mouse macrophages. We report that substitution of extracellular Cl− with non-halide anions, such as gluconate, results in a marked potentiation of caspase-1 activation and IL-1β processing. Cl− substitution also greatly potentiates the rapid release of mIL-1β to the extracellular compartment, and release of the cytokine is accompanied by an equally rapid release of caspase-1. We additionally demonstrate that ATP can stimulate release of IL-1β into the extracellular space by inducing an AG126-sensitive regulated secretion and a glycine-sensitive cytolysis, both of which can be independently manipulated. These results suggest that, under normal physiological conditions, P2X7R are maintained in a conformationally restrained state that coordinately limits ATP binding, channel gating, and coupling of the receptor to the protein complexes that regulate caspase-1 processing.
Materials and Methods
ATP and glycine were from Sigma-Aldrich. The tyrphostin tyrosine kinase inhibitor AG126 was purchased from Calbiochem. Abs for ELISA analysis of murine or human IL-1β were from Pierce. The monoclonal anti-IL-1β Ab 3ZD, which recognizes both 33 kDa pro-IL-1β and 17 kDa mIL-1β in Western blot analysis, was provided by the Biological Resources Branch, National Cancer Institute, Frederick Cancer Research and Development Center (Frederick, MD). Polyclonal anti-mouse caspase-1 Abs (sc-514 which recognizes p45 and p10), p38 MAPK Ab (sc-535), and HRP-coupled secondary Abs were from Santa Cruz Biotechnology. For analysis of LDH release, the Cytotoxicity Detection kit from Roche was used. LPS was purchased from List Biological Laboratories. All other salts and chemicals were of reagent grade. Newborn calf serum (iron-supplemented) was from HyClone Laboratories.
Macrophage models and cell culture
The BAC1.2F5 macrophage cell line, an SV40-transformed cell line derived from BALB/c murine macrophages, was generously provided by Dr. E. R. Stanley (Albert Einstein College of Medicine, New York, NY) and maintained as previously described in DMEM supplemented with 15% calf serum and 25% L cell-conditioned medium as a source of M-CSF (6). RAW264.7 cells, another murine macrophage line, was obtained from the American Type Culture Collection (ATCC) and maintained in DMEM supplemented with 10% calf serum. THP-1 human promonocytic leukemia cells from ATCC were cultured in RPMI 1640 medium supplemented with 10% calf serum and differentiated into adherent monocytes/macrophage-like cells by treatment with 10 nM PMA for 18–24 h before experiments. Bone marrow-derived macrophages (BMDM) from either BALB/c or C57BL/6 mice (Charles River Breeding Laboratories) were isolated based on a protocol modified from Davies and Gordon (35). Briefly, femurs and tibia were removed from the euthanized mouse and briefly sterilized in 70% ethanol. PBS was used to wash out the marrow cavity plugs and bone marrow cells were resuspended in L cell-conditioned medium containing M-CSF to stimulate proliferation and differentiation of the marrow progenitors into macrophages. After 7–9 days, the resulting BMDM were replated into six-well or 24-well plates and used within 2 wk. All experiments and procedures involving mice were approved by the Case Western Reserve University Institutional Animal Use and Care Committee.
Assays of caspase-1 activation and IL-1β processing and secretion
Murine macrophages or THP-1 human monocytes/macrophages were plated at 106/ml in 6-well plates (2 ml/well) for Western blot experiments, 12-well plates (1 ml/well) for K+ efflux experiments, or 24-well plates (0.5 ml/well) for ELISA experiments. For experiments involving LPS priming before stimulation with ATP, 1 μg/ml LPS in DMEM/10% calf serum/1% penicillin-streptomycin was added to the cells, followed by incubation at 37°C for 4 h. The culture medium was removed, cells were washed with PBS, and standard balanced saline solution (BSS) of NaCl-BSS containing 25 mM HEPES, 130 mM NaCl, 5 mM KCl, 1 mM CaCl2, 1 mM MgCl2, 5 mM glucose, and 0.01% BSA (pH 7.4) was used to bathe the cells for experiments. A total of 130 mM sodium-gluconate was substituted for the NaCl in standard BSS for experiments testing the effect of the removal of extracellular Cl−. In some experiments, the 130 mM NaCl in standard BSS was replaced with equimolar KCl, potassium gluconate, NaI, sodium glutamate, or sodium isethionate. For experiments using the pharmacological inhibitors (YVAD-cmk, AG126, glycine), the inhibitors were added directly to the various BSS bathing the macrophages for at least 15 min before stimulation with ATP and were present throughout the duration of the ATP pulse. Macrophages were stimulated with indicated concentrations of ATP (added directly from concentrated stocks), and after indicated time periods, the extracellular medium was collected and centrifuged briefly (to remove detached cells) for subsequent Western blot analysis of caspase-1, IL-1β, or p38 MAPK. Aliquots of the extracellular medium were also assayed for IL-1β by ELISA. The remaining adherent cells were lysed in buffer containing 25 mM HEPES (pH 7.7), 300 mM NaCl, 1.5 mM MgCl2, 0.2 mM EDTA (pH 8.0), 0.1% Triton X-100, 2 μg/ml leupeptin, 1 μg/ml aprotinin, 1 mM DTT, and 100 μg/ml PMSF.
Western blot analysis of cell-associated and extracellular caspase-1, IL-1β, and p38 MAPK
For assessment of cell-associated protein, cell lysates were separated by SDS-PAGE (15%) and electrophoretically transferred to polyvinylidene difluoride (PVDF) membranes under rapid-transfer conditions (in 1× Towbins buffer containing 3.03 g Tris, 14.4 g glycine, and 200 ml of methanol per liter) for 45 min at 1 mAmp. PVDF membranes were rinsed in immunoblot buffer (10 mM Tris (pH 7.4), 0.9% NaCl, 0.05% Tween 20, and 1 mM EDTA) and blocked for 1 h in immunoblot buffer containing 4% nonfat dried milk. PVDF membranes were then incubated in primary Ab in immunoblot buffer containing 4% milk at room temperature for 1 h. Primary Ab concentrations were as follows: 1/1000 for IL-1β, 1/200 for caspase-1, and 1/1000 for p38 MAPK. Membranes were then washed (1 × 15 min, 2 × 5 min) and incubated with 1/5000 dilutions of HRP-conjugated secondary Ab in 4% milk-immunoblot buffer for 1 h at room temperature. Membranes were washed and developed using chemiluminescent reagents (SuperSignal; Pierce) for 0.5–5.0 min and exposed to Eastman Kodak x-ray film. For routine assessment of proteins in extracellular media, the samples were concentrated by acid precipitation using 72 μl of 100% TCA and 15 μl of 10% cholic acid per 1 ml of extracellular media. The precipitated proteins were rinsed in acetone, followed by dissolution in Laemmli buffer. In some experiments involving analysis of IL-1β secretion, the extracellular media samples were incubated with biotinylated anti-mouse IL-1β (the same Ab used for ELISA) followed by precipitation with streptavidin-agarose and dissolution of the immune complexes in Laemmli buffer.
ELISA analysis of IL-1β release
From 1 to 25-μl aliquots of extracellular medium samples were assayed for IL-1β content by sandwich ELISA. Briefly, a 96-well plate was coated with 1 μg/ml primary anti-IL-1β overnight, then blocked with 4% BSA in PBS for 1 h. Plates were washed three times with wash buffer (50 mM Tris-HCl (pH 7.5), 0.2% Tween 20). Aliquots of medium samples or IL-1β standards (murine or human as applicable), diluted to 50 μl with PBS were added to the blocked wells along with 50 μl of the second biotinylated anti-IL-1β (murine or human as applicable) Ab (0.2 μg/ml). The plates were incubated at room temperature for 2 h and then washed three times. The captured immune complexes were colorimetrically detected by subsequent incubations with streptavidin-HRP conjugate (0.1 μg/ml), washing, development with tetramethyl benzidine substrate for HRP, and absorbance measurement using a Molecular Devices SoftMax Pro plate reader.
Measurement of K+ efflux
Bac1 macrophages were plated at 106/well in 12-well plates and primed for 4 h by incubation with 1 μg/ml LPS. The LPS priming medium was removed and each well was washed once with 1 ml of PBS before addition of 1 ml of either the NaCl or sodium gluconate BSS used for the IL-1β processing/release experiments. After a 10 min preincubation at 37°C, individual wells were supplemented with various concentrations of ATP (30–3000 μM) and incubated for an additional 10 min. Reactions were terminated by the rapid and complete aspiration of extracellular medium in each well. One milliliter of 10% nitric acid was added to each well to extract intracellular K+ from the adherent macrophages. After 3–4 h at room temperature, the nitric acid extracts were centrifuged and the supernatants assayed for K+ content by atomic absorbance flame spectrophotometry (PerkinElmer 3100).
Measurement of LDH release
Cells were cultured in 24-well dishes in 500 μl of medium per well. Upon completion of the experiment, the extracellular medium was removed and the cells were lysed in 500 μl of BSS containing 1% Triton X-100. Either 5–10 μl of cell lysate or 20–100 μl of extracellular medium was analyzed for LDH activity using the Cytotoxicity Detection kit from Roche. The results are expressed as a percentage of total LDH released, which was obtained by dividing the amount of LDH detected extracellularly by the sum total of LDH detected within the cell plus the amount detected in extracellular medium, and multiplying times 100.
Modulation of P2X7R-activated IL-1β release by replacement of extracellular Cl−
In the absence of ATP stimulation, LPS-primed Bac1 macrophages released only small amounts of IL-1β to the extracellular medium at a slow (<200 pg/106 cells/30 min) rate (Fig. 1,A). Activation of P2X7R with 1 mM ATP rapidly triggered a >40-fold increase in the rate of IL-1β release (8000 pg/106 cells/30 min) from cells incubated in a standard NaCl-BSS. Strong inhibition (90%) of the ATP-dependent IL-1β secretion was observed when the normal transmembrane gradients for Na+ and K+ were reduced by replacing the 130 mM extracellular NaCl with equimolar KCl or potassium gluconate (Fig. 1,A). However, these experiments also revealed that extracellular Cl− exerts a heretofore unappreciated inhibitory effect on the P2X7R→caspase-1→IL-1β signaling cascade because replacing the extracellular NaCl with 130 mM sodium gluconate resulted in further potentiation (by 5- to 10-fold) of the ATP-induced IL-1β release. Replacement of extracellular Cl− with various organic anions (glutamate, gluconate, or isethionate) facilitated this effect, albeit to different degrees (Fig. 1,B). In contrast, this potentiation was not produced upon substitution of the extracellular chloride with iodide (Fig. 1 A). Rather, this chaotropic halide strongly repressed the ATP-induced release of IL-1β.
Effect of Cl− replacement on the time course of P2X7R-activated IL-1β release
Time course experiments showed that the facilitative effect of Cl−/halide replacement could be observed at the earliest stages of P2X7R-dependent IL-1β release (Fig. 2,A). When Bac1 macrophages were activated by 1 mM ATP in either NaCl or sodium gluconate saline, an increased rate of IL-1β externalization was observed after a similar 5 min lag. Thereafter, the release rate was significantly amplified in cells incubated in sodium gluconate saline such that IL-1β externalization was maximally activated within 15 min. Although the Abs used for these ELISA-based measurements of IL-1β release have high avidity for the proteolytically processed form (17 kDa) of mIL-1β, they also react weakly with the 33 kDa pro-IL-1β precursor, which is the predominant intracellular form of this cytokine. To ensure that P2X7R activation in sodium gluconate was not simply enhancing cytolytic release of pro-IL-1β (31), we used an acid precipitation/Western blot protocol that permitted simultaneous analysis of pro-IL-1β and mIL-1β content within, and redistribution between the intracellular and extracellular compartments of Bac1 macrophages (Fig. 2,B). These experiments verified that mIL-1β was the predominant form of this cytokine in the extracellular media samples derived from cells stimulated in sodium gluconate test saline. The time course characterizing this enhanced release of mIL-1β was commensurate with that indicated by the ELISA-based assay of total IL-1β secretion (Fig. 2 A). Cells activated in NaCl-BSS also released significant amounts of mIL-1β, but at a rate significantly lower than was observed in the sodium gluconate BSS. Consistent with previous studies demonstrating a highly efficient mechanism for the externalization of IL-1β that has been processed by caspase-1, little intracellular accumulation of the mature cytokine was apparent during ATP stimulation of cells maintained in NaCl-BSS. In contrast, a substantial, but transient, intracellular accumulation of mIL-1β was observed during activation in the sodium gluconate medium. This observation suggested that replacement of Cl−/halide predominantly potentiated the caspase-1 mediated maturation of IL-1β as well as release of the processed cytokine.
Effect of Cl− replacement on the ATP concentration-response relationship for P2X7R-activated IL-1β release
Comparison of the concentration-response relationships characterizing ATP-dependent processing and release of IL-1β in NaCl vs sodium gluconate BSS indicated that replacement of extracellular Cl− with gluconate increased both the potency and efficacy of ATP (Figs. 3 and 4). Western blot analysis of the IL-1β released in NaCl-BSS revealed a biphasic action of ATP such that the relative proportions of unprocessed pro-IL-1β vs mIL-1β that was released varied as the ATP concentration was progressively increased (Fig. 3 A). Thus, although cells stimulated with 1 mM ATP predominantly released mIL-1β, cells stimulated with 3 mM ATP mainly released the unprocessed pro-IL-1β. These data suggest that Cl− influx through the nonselective pores induced by maximal activation of P2X7R (with high ATP) inhibits caspase-1 activity and thereby counteracts the activation of caspase-1 induced by K+ efflux through P2X7R cation channels. Replacement of the extracellular Cl− by gluconate eliminated this biphasic action of ATP such that mIL-1β was the predominantly secreted form even at the highest concentrations of ATP. Moreover, under these reduced Cl− test conditions, the release of mIL-1β was sharply induced as the ATP concentration is raised from 100 to 300 μM and the absolute amount of released mIL-1β was markedly increased at all tested ATP concentrations.
Cl− substitution also strongly enhanced the proteolytic processing and externalization of caspase-1 per se (Fig. 3 B). Although most of the 45-kDa procaspase-1 was present within the intracellular compartment of unstimulated Bac1 cells, a small but significant amount was constitutively released to the extracellular medium. When the macrophages were stimulated with ATP in NaCl-BSS, there was a time-dependent (data not shown) and dose-dependent increase in this redistribution of procaspase-1 from the intracellular to extracellular compartments. In the absence of ATP stimulation, only minor amounts of activated caspase-1 (as assayed by the content of 10-kDa subunit) were apparent in either the intracellular or extracellular pools. Significantly, production and externalization of activated caspase-1 in NaCl-BSS was maximal with 1 mM ATP. As with IL-1β, further elevation of ATP to 3 mM resulted in release of mainly unprocessed procaspase-1. In contrast, stimulation in sodium-gluconate saline elicited the predominant release of activated caspase-1 at all ATP concentrations. Under these conditions, the dose-response relationship for ATP-induced accumulation of the active 10-kDa subunit was extraordinarily steep over the 100 to 300 μM range. The amount of procaspase-1 released into the extracellular medium increased in an ATP dose-dependent and time-dependent manner in both NaCl and sodium gluconate BSS indicating that ATP may concurrently induce cytolysis.
The increase in ATP potency and efficacy by Cl− replacement was quantified in greater detail using the ELISA-based analysis that primarily detects mIL-1β (Fig. 4). Analysis of IL-1β released during activation of NaCl-BSS revealed a markedly biphasic concentration-response relationship (Fig. 4,A). The amount of ELISA-sensitive IL-1β progressively increased as the ATP stimulus was increased from 0.1 to 1 mM, but then dose dependently decreased as ATP was further increased from 1 to 3 mM. The low, overall potency of ATP as an activator of IL-1β processing and release (EC50 ∼800 μM) in NaCl-BSS was consistent with the low affinity for nucleotides that uniquely distinguishes P2X7R (31, 32). Replacement of the extracellular Cl− by gluconate eliminated the biphasic nature of the ATP concentration-response relationship and greatly increased the overall amount of released IL-1β (note the 10-fold difference in the y-axis scales of Fig. 4, A and B). As indicated by the normalized concentration-response relationships plotted in Fig. 4,C, replacement of Cl− with gluconate produced an ∼5-fold increase in the potency of ATP as an activator of IL-1β processing and release (EC50 ∼150 μM). This response is consistent with the pronounced effect of extracellular Cl− replacement on agonist potency at the P2X7R that we and others have previously described (33, 34). A similar 6-fold shift in agonist potency characterized the effect of Cl− replacement on ATP-stimulated K+ efflux (EC50 = 1.02 ± 0.28 mM in NaCl vs EC50 = 0.17 ± 0.04 mM in sodium gluconate, n = 3) as an independent index of P2X7R activity in these LPS-primed Bac1 macrophages (Fig. 4 D).
Effects of Cl− replacement on P2X7R-activated IL-1β release in other macrophage model systems
The effects of Cl− replacement on P2X7R-activated IL-1β processing and secretion were also evaluated in three other macrophage model systems: 1) primary mouse BMDM; 2) RAW264.7, which is another established murine macrophage cell line; and 3) THP-1 human promonocytic leukemia cells differentiated with phorbol ester into adherent monocyte/macrophage-like cells. Two allelic variants of the mouse P2X7R, reflecting single nucleotide polymorphism, have been identified (36). BALB/c mice (from which the Bac1 line was originally derived) express a Pro451 variant, whereas C57BL/6 (and other inbred strains) express a Leu451 P2X7R. Relative to the Pro451 P2X7R, the Leu451 variant is characterized by a significantly reduced efficacy of ATP in activation of cation influx, pore formation, and apoptotic induction in murine T lymphocytes (35, 36, 37); the effects of these allelic variants on P2X7R function in macrophages has not been tested thus far. Analysis of the concentration-response relationships for ATP-induced IL-1β release in BALB/c macrophages (Fig. 5,A) revealed both similarities and differences to the findings in Bac1 macrophages (Fig. 3,A). As in the Bac1 cells, supramillimolar ATP was required for significant processing and release of IL-1β in NaCl saline. However, increasing ATP to >1 mM did not promote release of 33-kDa pro-IL-1β but rather additional 17-kDa mature cytokine and an intermediate 22-kDa fragment. Significantly, stimulation of the BALB/c BMDM in sodium gluconate saline increased the potency and efficacy of ATP such that 600 μM was sufficient for maximal IL-1β secretion. Moreover, IL-β processing in these primary cells was very rapid and efficient such that most of the intracellular pro-IL-β was released as mIL-1β during the 30 min test incubation. Consistent with the reduced potency and efficacy of ATP as an agonist for the Leu451 P2X7R, a higher concentration of ATP (>1 mM) was required for maximal processing and release of IL-1β in C57BL/6 BMDM tested in sodium gluconate saline (Fig. 5 B). In NaCl medium, >3 mM ATP was required for optimal release of mIL-1β from these cells.
ATP stimulation of LPS-primed RAW264.7 murine macrophages in NaCl saline elicited markedly less release of IL-1β at all tested ATP concentrations as compared with Bac1 cells in paired experiments (Fig. 6,A). Although RAW264.7 macrophages in sodium gluconate saline showed increased IL-1β secretion in response to ATP ≥1 mM, little ELISA-detectable cytokine was released at lower ATP concentrations. Moreover, Western blot analysis of the extracellular medium from RAW64.7 cells stimulated with 3 mM ATP in sodium gluconate saline for 10–60 min revealed only unprocessed pro-IL-1β, which contrasted with the predominant secretion of mIL-1β in the time-paired samples from Bac1 cells subjected to the identical conditions (Fig. 6,B). Comparison of the whole cell lysates from these two murine macrophage lines indicated comparable expression of both procaspase-1 (Fig. 6,A, inset) and pro-IL-1β (Fig. 6 B). Thus, although Cl− replacement sensitizes the extent of ATP-stimulated pro-IL-1β release from RAW264.7 cells, this particular macrophage line appears to lack the signaling pathways that couple P2X7R-induced ionic fluxes to the caspase-1 activation machinery.
Replacement of extracellular Cl− with gluconate markedly increased the potency of ATP as a regulator of IL-1β secretion in THP-1 human monocytes/macrophages (Fig. 6 C). Although 3 mM ATP was required to induce significant IL-1β release in NaCl saline, submillimolar ATP was an effective stimulus in sodium gluconate medium. Thus, manipulation of extracellular Cl− also modulates regulation of the caspase-1 and IL-1β processing pathways by the human P2X7R.
Modulation of P2X7R-dependent caspase-1 activity by Cl− replacement
Fig. 7,A shows that the differential rate of IL-1β maturation during maximal P2X7R activation of Bac1 macrophages in NaCl vs sodium gluconate test media was not due to a simple collapse of the normal plasma membrane permeability barrier followed by proteolytic processing of pro-IL-1β in the extracellular compartment. Nonselective permeabilization of the cells with digitonin caused a release of pro-IL-1β into the extracellular compartment, but no processing of this released pro-IL-1β was apparent in either NaCl or sodium gluconate test media. This result indicated that the enhanced rate of IL-1β processing observed in sodium gluconate BSS reflects regulatory effects on intracellular caspase-1 activity in intact cells. This result was also supported by the finding that preincubation of the macrophages with YVAD-cmk, a membrane-permeable inhibitor of caspase-1, completely repressed the enhanced production of mIL-1β in sodium gluconate BSS resulting in release of only pro-IL-1β (Fig. 7 B).
ATP-induced cytolysis is glycine-sensitive and enhanced in sodium gluconate media
P2X7R activation has been shown to trigger a caspase-1 dependent cytolysis of macrophages, which is strongly potentiated by LPS priming (28). We tested whether Cl− replacement with gluconate increased the rate and extent of P2X7R-induced cytolysis by comparing ATP-induced LDH and IL-1β release from control or LPS-primed Bac1 macrophages stimulated in either NaCl or sodium gluconate media. Given the demonstrated ability of glycine to repress cytolysis in multiple cell types including macrophages (30), we additionally compared the effects of glycine on LDH release vs IL-1β secretion. Fig. 8,A shows that LPS priming potentiated ATP-induced LDH release from cells tested in either NaCl or sodium gluconate media and that cytolysis was greater in sodium gluconate saline (Fig. 8, A and C). However, this cytolysis was largely blocked by inclusion of 5 mM glycine in either test media. Glycine attenuated ATP-induced IL-1β release from the LPS-primed cells (Fig. 8, B and D) by <50% in both NaCl and sodium gluconate salines, indicating that much of the observed IL-1β secretion can be dissociated from cytolysis.
ATP-induced IL-1β secretion can be pharmacologically dissociated into regulated and cytolytic release mechanisms
We tested whether the enhanced cytolysis observed in sodium gluconate saline could be temporally or pharmacologically dissociated from secretion of mIL-1β. LPS-primed Bac1 macrophages in sodium gluconate BSS were subjected to a two-phase incubation protocol that involved initial stimulation with 500 μM ATP for a 15 min pulse, followed by removal of the ATP and secondary incubation in a fresh aliquot of ATP-free sodium gluconate BSS for varying periods of time. We used the tyrphostin inhibitor AG126 to block regulated IL-1β secretion (10, 20, 38) and glycine to block cytolysis (30). Parallel incubations were conducted in the continuous presence of 50 μM AG126 alone, 5 mM glycine alone, both AG126 and glycine, or in the absence of both inhibitors. Extracellular media samples from both the primary ATP pulse incubation and the secondary agonist-free incubation were analyzed by Western blot for IL-1β, caspase-1, and p38 MAPK (as a marker of nonselective cytolysis). Fig. 9,A shows that, in the absence of AG126 or glycine, the 15 min ATP test pulse elicited secretion of mIL-1β and active caspase-1. Secondary incubation of these cells for increasing times in fresh medium lacking ATP revealed a delayed phase of cytolysis characterized by the extracellular accumulation of pro-IL-1β, procaspase-1, and p38 MAPK. Incubation of the macrophages in medium containing 5 mM glycine blocked this delayed lytic release of pro-IL-1β, procaspase-1, or p38 MAPK, but not the release of mIL-1β and activated caspase-1 induced during the initial 15 min ATP test pulse. Conversely, inclusion of 50 μM AG126 completely repressed release of mIL-1β and active caspase-1 during the primary incubation, but not the secondary cytolytic release of pro-IL-1β, procaspase-1, and p38 MAPK. Finally, coincubation with both glycine and AG126 led to the complete blockade of ATP-induced release of IL-1β, caspase-1, or p38 MAPK at all time points examined. The anticytolytic action of glycine but not AG126 under these conditions was further confirmed by LDH assays (Fig. 9,B). ELISA revealed significant and AG126-sensitive IL-1β secretion predominantly during the initial ATP pulse incubation (Fig. 9 C).
These studies indicate that replacement of extracellular Cl− with non-halide, nonchaotropic anions strongly potentiates the coupling of P2X7R to the IL-1β processing and secretion machinery in macrophages. Our findings confirm and extend previous reports that investigated effects of extracellular anions on various P2X7R functions in immune and inflammatory cells. Perregaux et al. (39) were the first to demonstrate that substitution of extracellular Cl− with iodide or chaotropic organic anions, such as thiocyanate, repressed P2X7R activation of IL-1β processing and release in human monocytes. Those authors also noted that substitution with glucuronate did not mimic this inhibitory effect and actually caused a modest increase in ATP-induced IL-1β secretion. More recently, Marty et al. (40) reported that replacement of extracellular Cl− with aspartate enhanced ATP-induced IL-1β processing and release from LPS-primed Schwann cells, whereas substitution with thiocyanate attenuated this response. Finally, Tsukimoto et al. (41) have described a critical involvement of Cl− influx in the ability of ATP to trigger apoptotic death of chicken DT40 B lymphocytes stably transfected with rat P2X7R. Our specific observations regarding multiple phases of the P2X7R→caspase-1 activation→IL-1β secretion cascade strongly suggest that these diverse effects of extracellular Cl− and Cl− substitution involve three levels of regulation, which may vary in relative contribution depending on cell background. Our data also provide further evidence that the P2X7R-induced release of IL-1β can be unequivocally dissociated from the cytolysis that generally accompanies sustained P2X7R activation in leukocytes.
A primary mechanism for this regulatory action of extracellular Cl− is consistent with an apparent negative allosteric action at the level of the P2X7R itself. This will limit the ability of ATP to gate P2X7R conformations required for channel activation and efficient coupling to the IL-1β processing and secretion machinery. Voltage-clamp analysis of recombinant human P2X7R indicated that these channels are strongly modulated by Cl− removal as reflected in a 22-fold increase in the potency of nucleotide agonists in the gating of inward Na+ current (34). Similarly, substitution of extracellular Cl− with non-halide anions increases the potency and efficacy of ATP as an agonist for P2X7R-mediated K+ efflux (Fig. 4 D) and macropore induction (33). These observations indicate that different extracellular anions allosterically modulate P2X7R function either by directly affecting the binding of the ATP agonist or indirectly perturbing the active conformational states induced by agonist binding. Very recent studies by Li et al. (42) have indicated a role for extracellular Cl− in modulating the permeability changes triggered by P2X7R in murine parotid acinar duct cells. Like other P2X family receptors, the P2X7R channel is likely assembled as a stable trimer of individual P2X7 subunits (43). Manipulation of extracellular Cl− may allosterically modify the PX7R trimers with consequent changes in the packing of the extracellular loops from the individual subunits.
Although gating of cation channel function is the most immediate consequence of ligand-induced changes in P2X7R conformation, increasing evidence suggests that P2X7R additionally act as docking sites for a variety of intracellular proteins. Surprenant and colleagues (44) have used proteomic analysis of anti-P2X7R immunoprecipitates to demonstrate P2X7R association with multiple signaling and cytoskeletal proteins. All P2X receptor subunits share a similar structure consisting of intracellular N and C termini, two transmembrane segments, and a large extracellular loop containing 10 similarly spaced cysteines and glycosylation sites (45). The P2X7R is significantly larger (595 amino acids) than the other six P2X subtypes due to its intracellular C terminus, which is two to eight times longer than the C termini of the other receptors. Recent studies have demonstrated this intracellular C-terminal tail contains the molecular determinants for the regulated trafficking of the P2X7R to the plasma membrane as well as its ability to induce the nonselective macropore and membrane blebbing (46, 47, 48). Thus, the P2X7R C-terminal domain may play a role as an oligomeric scaffold for the initial assembly/activation of adapter proteins that interact with procaspases and which activate other membrane proteins. If so, allosteric perturbation of P2X7R trimers by extracellular Cl− or Cl− removal could also affect the packing of adjacent intracellular C termini and thereby affect the avidity or efficiency of associated protein interactions.
An additional mechanism by which perturbation of extracellular Cl− modulates P2X7R regulation of IL-1β secretion appears to involve intracellular accumulation of Cl− via the nonselective pores induced by activated P2X7R. Stimulation of the P2X7R rapidly induces a nonselective pore that facilitates flux of both cationic and anionic molecules up to 800 Da in mass. Sustained activation of these pores in standard Cl− containing salines can lead to a progressive elevation of intracellular Cl− concentration due to: 1) the chemical gradient that normally favors Cl− influx in most cells (130 mM extracellular Cl− vs 10–40 mM intracellular Cl−); and 2) the P2X7R-mediated collapse of the negative membrane potential that otherwise limits Cl− influx in the resting state. By assaying apical Ca2+-dependent Cl− fluxes in epithelial cells engineered to express basolateral P2X7R, Gabriel et al. (49) provided direct evidence for P2X7R-induced Cl− accumulation. This progressive Cl− accumulation would obviously be eliminated when P2X7R are activated in sodium gluconate test media. Rather, such cells will contain the nonchaotropic gluconate as the major osmotically active anion. Excessive accumulation of Cl− rather than organic anions may attenuate or alter the intracellular assembly of the signaling complexes required for caspase-1 activation and IL-1β processing. In this regard, much recent research has focused on the molecular platforms known as “inflammasomes,” which facilitate the activation of procaspase-1 (12, 13). Inflammasome complexes involve direct CARD-CARD binding between ASC and procaspase-1 and PYRIN-PYRIN interactions between NALP family proteins and ASC or additional adapters. These interactions bring procaspase-1 monomers and/or other inflammatory procaspases (human caspase-5 or murine caspase-11) into induced proximity, which facilitates their autoproteolytic activation. Martinon et al. (12) initially characterized the spontaneous assembly of inflammasome complexes in cell-free lysates of THP-1 monocytes that were disrupted and assayed in a hypotonic, low-salt buffer. Researchers speculated that physical disruption of the cells induced the generation or release of a regulatory ligand that allosterically interacts with the NALP family adapters that initiate inflammasome assembly; this results in increased processing of caspase-1 and/or IL-1β processing in the cell-free lysates. By adapting similar assay conditions with Bac1 macrophages, we recently reported that brief stimulation of P2X7R immediately before cell disruption greatly increased the rate of caspase-1 processing in the resulting cell-free lystates (10). Using this system we have found that inclusion of 25 mM Cl−, but not 25 mM gluconate, in the lysate buffer prevents this rapid caspase-1 processing (data not shown). This finding supports the possibility that elevated intracellular Cl− can act to attenuate inflammasome formation.
It should be noted that manipulation of other extracellular ions exerts complex effects on both general P2X7R function and its ability to regulate the IL-1β processing pathway. Extracellular Na+ has been shown to repress P2X7R-induced ethidium influx in human lymphocytes (50) and to modulate nucleotide-induced N-methyl-d-glucamine currents in cells expressing recombinant P2X7R (42, 51). However, Perregaux and Gabel (52) demonstrated that physiological levels of extracellular Na+ are required for optimal activation of IL-1β processing and secretion in human monocytes stimulated with either ATP or nigericin, a monovalent cation ionophore that induces a P2X7R-independent activation of caspase-1. Thus, extracellular Na+, like extracellular Cl−, may modulate ATP-induced IL-1β processing and secretion at both P2X7R activation per se and signaling steps downstream of the activated receptor. The ability of nucleotide agonists to activate P2X7R is also strongly affected by the extracellular concentrations of the divalent cations, Mg2+ and Ca2+, such that the potency of ATP is enhanced by a reduction in Mg2+ and/or Ca2+ (31, 32). Because organic anions, such as gluconate, have a higher affinity than Cl− for divalent cations, it is possible that the observed left-shifts in the concentration-response relationships describing ATP activation of IL-1β secretion (Figs. 4 and 5) additionally involves perturbation of local divalent cation levels.
Importantly, we demonstrated that the results obtained in the Bac1 macrophage cell line can be extended to primary mouse macrophages, with some subtle and interesting differences. We used macrophages from two different mouse strains (BALB/c and C57BL/6) to illustrate the importance of polymorphic alleles of the P2X7R, and because the Bac1 macrophage cell line was originally derived from SV40-transformed peritoneal macrophages from a BALB/c mouse. The most striking difference between the cell line and the primary BALB/c macrophages was that the latter cells were much more efficient in converting and releasing intracellular IL-1β pools as mIL-1β (compare Figs. 3,C and 5). This finding may indicate that macrophages at different stages of development or activation are characterized by: 1) differential expression of various inflammasome components; 2) distinct rates of ATP-induced cytolysis; or 3) altered extents of nonselective pore formation. This indication is supported by our finding that RAW264.7 cells, another widely used macrophage cell line, mainly released pro-IL-1β in response to ATP even when stimulated in sodium gluconate saline (Fig. 6,B). The higher threshold concentrations of ATP required to elicit cytokine release from the BMDM (both in NaCl and sodium gluconate BSS) may be due to lower cell surface expression of P2X7R channels in primary macrophages. All of these possibilities are currently being evaluated. Finally, we observed that replacement of extracellular Cl− with gluconate induced a left-shift in the dose-response curve for ATP activation of IL-1β secretion from THP-1 monocytes/macrophages (Fig. 6 C). Thus, extracellular Cl− similarly modulates the P2X7R→caspase-1 activation cascade in human inflammatory leukocytes.
The reduced potency of ATP as an activator of IL-1β release from C57BL/6 cells (relative to BALB/c macrophages) underscores the likely role of P2X7R polymorphisms in enhancing or limiting the actions of P2X7R as a mediator of innate immunity in different mouse strains. Significantly, similar polymorphic variations of the human P2X7R have been shown to attenuate ATP-induced secretion of IL-1β and IL-18 from monocytes isolated from different human subjects (53, 54).
This study also clarifies the relative contributions of a glycine-sensitive cytolytic pathway vs regulated nonlytic mechanisms to ATP-induced IL-1β release in various experimental contexts. We confirmed the recent observations of Le Feuvre et al. (28), which indicated that LPS primes macrophages for ATP-dependent cytolysis. Our finding that removal of extracellular Cl− markedly potentiated cytolysis in parallel with its effects on IL-1β release is consistent with a role for caspase-1 in ATP-induced cytolysis as also proposed by Le Feuvre et al. (28). Interestingly, infection of macrophages with Salmonella or Shigella also induces a coordinated activation of caspase-1 followed by rapid cell death (55, 56, 57). However, the ability of glycine to greatly reduce ATP-elicited cytolysis while only partially diminishing IL-1β release indicates that the regulated secretion of mIL-1β can be dissociated from rapid macrophage death. Indeed, experiments using a relatively brief (15 min) pulse stimulation with ATP, followed by removal of the ATP stimulus and extended incubation, revealed that most of the mIL-1β was rapidly released by a nonlytic mechanism. This mechanism likely involves rapid shedding of evaginated microvesicles (24) and/or exocytosis of exosomes packaged within specialized secretory lysosomes (25) before the delayed onset of cytolysis. The mechanism by which glycine confers protection against ATP-induced cytotoxicity is unknown but it is significant to note that glycine produces a similar blockade of the macrophage death induced by Salmonella or Shigella infection (55, 57).
These results suggest that under physiological conditions, i.e., in the presence of high extracellular Cl−, the P2X7R is maintained in a conformationally constrained state that coordinately limits ATP binding, channel gating, and coupling of the receptor to the protein complexes that regulate caspase-1 processing. In the artificial context of the tissue culture environment, reduction in the extracellular Cl− concentration can relieve this constrained conformation. However, with the exception of certain specialized tissue spaces such as apical interfaces of the airway and digestive tract (42), Cl− is steadily maintained at high levels in most relevant extracellular compartments in vivo. Thus, factors or environmental conditions other than reduced Cl− are required to induce similar changes in receptor conformation within the monocytes/macrophages that populate inflammatory loci. Identifying these factors or conditions will require innovative approaches for analyzing P2X7R activation and function in situ.
The authors have no financial conflict of interest.
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.
This work was supported by National Institute of General Medical Sciences Grant GM-36387 (to G.R.D.) and partially supported by Medical Scientist Training Program Grant T32-GM07250 from the National Institutes of Health (to P.A.V. and J.M.K.).
Abbreviations used in this paper: ROCK, Rho-effector kinase; BMDM, bone marrow-derived macrophage; mIL-1β, mature IL-1β; PVDF, polyvinylidene difluoride; LDH, lactate dehydrogenase; BSS, balanced saline solution.