Signaling by extracellular nucleotides through P2 purinergic receptors affects diverse macrophage functions; however, its role in regulating antimicrobial radicals during bacterial infection has not been investigated. Mycobacterium tuberculosis-infected macrophages released ATP in a dose-dependent manner, which correlated with nitrite accumulation. P2 receptor inhibitors, including oxidized ATP, blocked NO synthase (NOSII) up-regulation and NO production induced by infection with M. tuberculosis or bacille Calmette-Guérin, or treatment with LPS or TNF-α. Oxidized ATP also inhibited oxygen radical production and activation of NF-κB and AP-1 in response to infection and inhibited NO-dependent killing of bacille Calmette-Guérin by macrophages. Experiments using macrophages derived from P2X7 gene-disrupted mice ruled out an essential role for P2X7 in NOSII regulation. These data demonstrate that P2 receptors regulate macrophage activation in response to bacteria and proinflammatory stimuli, and suggest that extracellular nucleotides released from infected macrophages may enhance production of oxygen radicals and NO at sites of infection.
While reactive nitrogen and oxygen species play a vital role in microbial killing, they are also important mediators of inflammation and tissue damage (1, 2, 3); thus, their expression is tightly regulated. High output production of reactive oxygen intermediates (ROI)5 (3) and reactive nitrogen species such as NO generally requires two signals: “priming” with IFN-γ, and a second “activating” stimulus such as bacterial infection (1, 2, 3). However, other regulatory mechanisms may further limit NO and ROI production to appropriate sites of action.
Signaling by extracellular nucleotides through P2 purinergic receptors represents one such potential mechanism. The P2 family of receptors comprises several subfamilies of heterogeneous structure and function (P2U, P2X, and P2Y) (4, 5) that are stimulated by nucleotide tri- and diphosphates, including ATP. Macrophages express diverse P2 receptor subtypes (6), and extracellular nucleotides have been reported to affect numerous macrophage activities (7, 8, 9, 10, 11, 12, 13, 14). In light of the vital roles played by oxygen and nitrogen radicals in host defense, experiments were performed to test the hypothesis that signaling through P2 receptors modulates the production of NO and ROI by murine macrophages during infection with bacteria.
Materials and Methods
Reagents, cells, animals, and bacteria
8-(p-Sulfophenyl)theophylline (SPX) was obtained from RBI (Natick, MA); pyridoxalphosphate-6-azophenyl-2′,4′-disulfonic acid (PPADS), l-n-monomethyl arginine (l-NMMA), and TNF-α from Calbiochem (La Jolla, CA); recombinant murine IFN-γ from Boehringer-Mannheim (Indianapolis, IN); and bacterial media from Difco (Detroit, MI). All other reagents are from Sigma (St. Louis, MO). The plasmid pIg κ-Luc has been described previously (15). The plasmid 6AP-1-luc was kindly provided by Dr. Roya Khosravi-Far (Department of Biology, Massachusetts Institute of Technology). The control plasmid pCMVB, which contains the lacZ gene under control of the constitutive CMV immediate early promoter, was obtained from Clontech (Palo Alto, CA).
The murine macrophage-like cell line J774 was maintained in complete DMEM as described (16). Primary bone marrow-derived macrophages (BMDM) were obtained from C57BL/6 mice (The Jackson Laboratory, Bar Harbor, Maine) or P2X7−/− mice (Glaxo Wellcome, U.K.; as described in G.B., I.C., and K. Duncan, manuscript in preparation) (17). For experiments, pretitered frozen suspensions of BCG or Mycobacterium tuberculosis were thawed at room temperature, diluted 1:10 into DMEM, and sonicated briefly with a probe sonicator to disperse clumps. The number of Escherichia coli organisms per milliliter of saturated culture of strain DH5α (Life Technologies, Grand Island, NY) was determined with a Petroff-Hausser counting chamber (VWR Scientific, West Chester, PA).
Cell culture assays for NO production, ATP release, and NO-dependent killing of bacteria
J774 cells or BMDM were plated into 96- or 12-well tissue culture plates, cultured for 16 h in the presence of complete DMEM containing 100 U/ml recombinant murine IFN-γ, washed, and treated with the indicated inhibitors and NO-inducing stimuli. Except where noted, bacille Calmette-Guérin (BCG) was added at a multiplicity of 30 viable organisms per macrophage; LPS was used at 1.0 μg/ml; and TNF-α at 100 ng/ml. oATP was added for a 2-h preincubation at 300 μM and then washed out. NO production was determined after 16 h by assaying nitrite in culture supernatants by the Greiss reaction (18). ATP measurements were made using a luciferin/luciferase assay system (Sigma). BCG-killing experiments were done as described (19).
Transient transfection and reporter assays
J774 cells were cotransfected with the pIg κ-Luc or 6AP-1-luc constructs and the p-CMVB control plasmid using Lipofectamine Plus (Life Technologies). For reporter assays, cells were primed with IFN-γ, incubated in the presence or absence of 300 μM oATP for 2 h, washed, and stimulated for 5 h with LPS (2 μg/ml) or BCG (30 CFU/cell). Cell lysates were then assayed for luciferase and lacZ activity using commercial reporter assay kits (Promega, Madison, WI).
Assays of oxidative burst
J774 cells or BMDM were detached, washed, and preincubated in the absence or presence of 300 μM oATP for 2 h, and in the absence or presence of 1 mM l-NMMA for 1 h. Cells were then washed and resuspended in antibiotic-free DMEM in the absence or presence of 1 mM l-NMMA (to control for the contribution of nitrogen radicals to luminol chemiluminescence), at a concentration of 106 cells/ml. For each assay, an aliquot of 105 cells was stimulated with 1 μg/ml PMA for 10 min or with 100 E. coli per cell for 1 h. Production of ROI was measured at the indicated time points by assaying aliquots of 105 cells for luminol-enhanced chemiluminescence (20).
NO synthase immunostaining and Northern blot analysis
NO synthase (NOSII) immunostaining was performed on paraformaldehyde-fixed cell monolayers using a mouse NOSII-reactive IgG1 primary Ab obtained from Transduction Laboratories (Lexington, KY; cat. No. N39120), and developed using the Vectastain Elite ABC kit (Vector; Burlingame, CA). For Northern blot analysis, J774 macrophages were grown to confluence in 6-well tissue culture plates and treated with oATP and BCG or LPS as described above. After 8 h, RNA was isolated with Trizol (Life Technologies). For each condition, 20 μg were subjected to Northern blot analysis (21), using a 32P-labeled cDNA probe derived from the rat NOSII gene (kindly provided by Dr. Douglas Feinstein, Department of Neurology and Neuroscience, Cornell University Medical School). Blots were then stripped and reprobed for GAPDH.
Results and Discussion
P2 receptor inhibitors block NO production and NO-mediated bacterial killing
A panel of chemical inhibitors (suramin, oATP, and PPADS) was used to test the hypothesis that P2 receptors control BCG-induced NO production by murine macrophages. All three P2 receptor antagonists inhibited BCG-induced NO production by J774 cells (Fig. 1 A). In contrast, an inhibitor specific for P1 (adenosine) receptors, SPX, did not affect NO production. None of the inhibitors were cytotoxic, as measured by trypan blue exclusion, lactate dehydrogenase release, and activity of the mitochondrial enzyme MTT (data not shown). Since oATP inhibition is irreversible (22), it was used for all additional experiments to avoid direct effects of the inhibitor upon bacteria. ATP alone, at concentrations from 1 nM to 10 mM, did not itself induce NO release (data not shown), suggesting that P2 receptor stimulation is a necessary but not a sufficient stimulus for the induction of NO.
Additional experiments investigated whether P2 receptors control NO production in response to other stimuli. oATP pretreatment of J774 cells blocked nitrite production induced by LPS or TNF-α (Fig. 1 B). oATP also inhibited nitrite production by primary murine macrophages in response to TNF-α or infection with virulent M. tuberculosis, demonstrating that inhibition is not a cell line-specific effect. These observations are fully consistent with data suggesting a role for P2 receptors in the modulation of macrophage NO production in response to LPS (8, 9, 10, 23), and with recent findings that NO production and NF-κB activation in a macrophage-like cell line (24) and in human astrocytes5 is inhibited by P2 antagonists.
The P2X7 purinergic receptor (25) has been linked to diverse macrophage activities, including apoptosis (11), release of the proinflammatory cytokine IL-1β (12), and NO-independent killing of intracellular mycobacteria (13, 14). The recent availability of P2X7 knockout mice allowed us to determine whether the P2X7 receptor might also be involved in the regulation of NO production. (Fig. 1 C). When stimulated with LPS or BCG, P2X7−/− BMDM made NO in quantities comparable to wild-type BMDM and were sensitive to oATP inhibition. This formally excludes an essential role for P2X7 in the regulation of NO production by BMDM and suggests that one or more other P2 receptors are required.
Murine macrophages, when activated with IFN-γ and LPS in vitro, can kill M. tuberculosis or BCG through an NO-dependent mechanism (19). To determine whether inhibition of P2 receptors impaired the macrophage’s ability to kill bacteria, we examined the effect of oATP on killing of intracellular BCG by IFN-γ/LPS-activated J774 cells (Fig. 1 D). Activated macrophages produced NO and reduced the viability of intracellular BCG by 58% over a 24-h incubation. Pretreatment with oATP did not alter the efficiency of infection (not shown) but did reduce NO production and inhibit bacterial killing by over 65%.
Regulation of NOSII expression in response to BCG and LPS
Since high output NO production in macrophages is catalyzed by NOSII, experiments were performed to determine whether oATP blocked NO production by inhibiting the accumulation of NOSII protein and mRNA. NOSII immunostaining demonstrated that preincubation with oATP blocked NOSII protein induction by BCG-infected J774 cells (Fig. 2,A). Northern blot analysis demonstrated that up-regulation of NOSII mRNA in BCG-infected J774 cells was abolished by pretreatment with oATP (Fig. 2B). These data demonstrate that oATP blocks NO production by inhibiting the accumulation of NOSII mRNA and protein.
Since NF-κB and AP-1 binding sites are present in the NOSII promoter and involved in regulation of NOSII (1, 26), we performed experiments to determine whether oATP inhibits NOSII transcription by interfering with activation of NF-κB- and AP-1-dependent gene transcription. Pretreatment with oATP blocked induction of TNF-α-, LPS-, or BCG-induced NF-κB activity by 82%, 77%, or 65%, respectively (Fig. 3,A). However, oATP did not inhibit H2O2-induced NF-κB activity (Fig. 3,B), suggesting that oATP inhibition is specific for receptor-mediated pathways of NF-κB activation. Similar results were seen for BCG-induced activation of AP-1 (Fig. 3 C). Based on these data, we infer that inhibition of P2 receptors blocks NOSII synthesis by interfering with activation of NF-κB and/or AP-1. Such a role for P2 receptors in the regulation of NF-κB activation is consistent with recent observations that high (mM) concentrations of ATP and ADP can themselves trigger NF-κB activation (27, 28).
oATP inhibits ROI production by activated macrophages
Like NO, microbicidal ROI are produced by activated macrophages and play an important role in host defense (3). We examined the effect of oATP pretreatment on the production of ROI by J774 cells in response to PMA or to infection with E. coli (Fig. 4). Pretreatment with oATP inhibited ROI production in response to either stimuli. In addition, P2X7−/− bone marrow-derived macrophages produced ROI in quantities similar to those produced by wild-type macrophages in response to PMA treatment (data not shown). We conclude that P2 receptor blockade inhibits ROI production, probably through a P2X7-independent mechanism.
The present results suggest that classical macrophage-activating stimuli (bacterial infection, LPS, and PMA) synergize with P2 receptor-mediated signals to induce production of NO and ROI, and demonstrate a previously unsuspected role for purinergic receptors in the regulation of antimicrobial effector functions. Extracellular nucleotides are likely to be present at sites of bacterial infection; we observed that infection with M. tuberculosis caused macrophages to release ATP into the culture supernatant in a dose-dependent fashion, correlated with the increase in nitrite accumulation (Fig. 5). This probably reflects the liberation of intracellular nucleotides from cells lysed by M. tuberculosis. The importance of extracellular nucleotides in the macrophage response to mycobacterial infection has been confirmed by recent data demonstrating that treatment with P2 inhibitors or the ATP-degrading enzyme apyrase blocks macrophage Ca2+ signaling in response to BCG infection (A. S., S. Suadicani, B. R. B., and D. Spray, manuscript in preparation). In vivo, release of intracellular nucleotides from platelets and dead and injured cells leads to high levels of extracellular nucleotides at sites of inflammation and tissue damage (4, 5, 6). Since ROI and NO may themselves contribute to pathological inflammation and toxicity, regulating their production through extracellular nucleotides may promote expression at sites of infection while sparing unaffected tissue, thus limiting potentially harmful macrophage effector functions to appropriate physiological contexts.
We are grateful for helpful suggestions from Jeffery S. Cox, Mary K. Hondalus, and Michael L. Maitland.
Data in this paper are from a thesis to be submitted in partial fulfillment of the requirements for the Degree of Doctor of Philosophy in the Sue Golding Graduate Division of Medical Sciences, Albert Einstein College of Medicine, Yeshiva University.
Abbreviations used in this paper: ROI, reactive oxygen intermediates; BCG, bacille Calmette-Guérin; BMDM, bone marrow-derived macrophages; l-NMMA, l-n-monomethyl arginine; oATP, oxidized ATP; PPADS, pyridoxalphosphate-6-azophenyl-2′,4′-disulfonic acid; SPX, 8-(p-sulfophenyl)theophylline; NOSII, NO synthase.
J. S. H. Liu, G. John, A. Sikora, S. C. Lee, and C. F. Brosnan. Nuclear factor κB activation by interleukin-1 by human fetal astrocytes is regulated by purinergic receptors. Submitted for publication.