To determine whether ATP was capable of inducing autophagy in THP1 cells and monocyte-derived macrophages (MDMs), the expression of the microtubule-associated (LC3-I) and membrane associated (LC3-II) forms of LC3 protein were assessed in both ATP-treated and untreated cells by Western Blot and Confocal microscopy.

The processing of the 18 kDa microtubule-associated form of LC3 protein (LC3-I) to its 16 kDa, truncated form (LC3-II), is accepted as one of the standard read-outs of cell autophagy 910. In Western blot studies, we observed expression of both LC3-I and LC3-II forms of LC3 protein in ATP-treated THP1 cells (Figure 1A) and MDMs, (Figure 1B) as revealed by bands with electrophoretic mobility corresponding to molecular masses of 18 kDa and 16 kDa respectively. In contrast, only the 18 kDa LC3-I band was expressed in control non-ATP treated THP1 or MDM cells (Figures 1A and 1B).

Another commonly used assay for monitoring for the occurrence of cell autophagy relies on demonstrating by microscopy the transitional shift of the 18 kDa cytosolic (microtubule-associated) form of LC3 (LC3-1) to its membrane-associated, 16 kDa lipidated form, LC3-II on autophagosomes 910. Using an anti-LC3 antibody, we observed, by confocal microscopy, an increased transition of the diffuse cytosolic staining pattern, typical of LC3-I, (Figure 2A; left panel) to the punctuate staining pattern, characteristic of LC3-II, within ATP-treated MDMs (Figure 2A; right panel).

We have previously reported that ATP can stimulate Ca²⁺ entry into cells not only through activation of ionotropic P2X purinoceptors, but also via mobilisation of intracellular Ca²⁺ into the cytosol by activation of metabotropic P2Y receptors 7. We then sought to establish the relative contribution of these Ca²⁺ sources to the induction of autophagy by examining the cellular response to ATP (3 mM) stimulation for 30 minutes in both Ca²⁺-containing and Ca²⁺-free medium. In cells treated with ATP in Ca²⁺-containing media we observed the induction of autophagy by Western blot analysis as revealed by the expression of both the 16 kDa and 18 kDa forms of LC3 protein (Figure 1C). In contrast, ATP-induced cell autophagy was shown to be completely inhibited in Ca²⁺-free medium, as revealed by the absence of expression of the 16 kDa lipidated form of LC3 (LC3-II) (Figure 1C).

Having obtained preliminary evidence of P2X receptor involvement in ATP-induced cell autophagy from the above results, we sought evidence that this process involved P2X₇ receptors. We then investigated the induction/inhibition of autophagy in response to a number of purinergic receptor agonists and antagonists. It was observed in Western blot assays (Figure 1D) that autophagy (LC3-II expression) was not induced following uridine 5'-triphosphate (UTP) (3 mM) stimulation, which activates P2Y, but not P2X receptors. In contrast, autophagy was induced by the ATP analogue, benzoylbenzoic ATP (bzATP), which selectively activates P2X₇ receptors. To further confirm the role of the latter receptor, we pre-treated THP1 and MDM cells with either adenosine 5'- triphosphate periodate oxidized ATP (oATP); an irreversible antagonist of P2X₇, or an anti-P2X₇ blocking antibody, before treating with ATP. Autophagy was completely blocked in the presence of either inhibitor, as demonstrated by the absence of LC3-II expression in Western blot gels (Figure 1D), thereby confirming that ATP-induced autophagy is mediated via activation of P2X₇ receptors.

We have previously reported that ATP stimulation of P2X₇ promotes acidification of BCG-containing phagosomes within infected human macrophages 7. Moreover, it has also been reported that induction of autophagy in RAW cells by starvation or treatment with rapamycin was associated with acidification and maturation of mycobacterial phagosomes 4. In view of these findings, we analyzed the role of autophagy in ATP-mediated phago-lysosomal fusion within MDMs infected with green fluorescent protein expressing M. bovis BCG (GFP-BCG). Lysosomal trafficking was monitored by preloading BCG-infected cells for 1 hour with the acidotropic fluorescent dye Lysotracker Red that accumulates within acidic organelles 11. The cells were then stimulated with ATP (3 mM) for 30 minutes. While control non-ATP treated cells showed minimal co-localisation of GFP-BCG with Lysotracker Red labeled lysosomes (7.2 ± 2.5%), there was significantly greater co-localisation within ATP-treated cells (63.3 ± 4.5%) (p < 0.01; student's unpaired t test) (Figures 2B &2C). The results confirmed our previously reported findings on the ability of ATP to induce phago-lysosomal fusion within BCG-infected MDMs 7. We next examined whether pharmacological inhibitors of autophagy could block this effect. We observed that wortmannin, a classical phospatidylinositol-3 kinase (PI3 Kinase) inhibitor, which has previously been reported to inhibit the early stages of autophagosome formation 12, dramatically reduced the levels of co-localisation of BCG-containing phagosomes with lysosomes (29.5 ± 3.8%) (p < 0.01; student's unpaired t test) (Figures 2B &2C). The results support the involvement of cell autophagy in ATP-induced, phagolysosome fusion and acidification of mycobacteria-containing phagosomes. Pre-treatment of MDMs with an anti-P2X₇ neutralising antibody also resulted in a significant reduction in the levels of co-localisation (36.8 ± 5.4%) (p < 0.01; student's unpaired t test) (Figures 2B &2C).

Electron microscopic analysis of ATP-treated MDMs revealed the presence of multiple, large cytosolic autophagic vacuoles within both non-infected and BCG-infected cells (Figures 3B and 3C), while such vacuoles were almost absent in non- ATP-treated, infected cells (Figure 3A). In addition within ATP-treated infected cells, EM analysis revealed the presence of bacilli within characteristic double-membrane autophagosomes containing partially degraded internal membranes typical of maturing autophagosomes (Figure 3C &3D) 121314.

We have previously reported that ATP induces rapid killing of intracellular mycobacteria via a cell-mediated process, as ATP had no direct bactericidal effect on BCG 6. To examine the role of autophagy in this process, BCG-infected MDMs were pre-treated with wortmannin for 1 hr prior to pulsing with ATP (3 mM) for 30 minutes and the viability of the intracellular mycobacteria assessed at 2, 4, and 6 hours post-treatment by ³H-uridinine incorporation assay (Figure 4). ATP was shown to result in a marked reduction in intracellular BCG viability at 2 hr (58.6 +/- 6.35%, p < 0.05), 4 hr (45.1 ± 3.7%, P > 0.005) and 6 hrs post treatment (31.9 +/- 6.2% p < 0.005) as compared to that obtained in non-ATP-treated, infected cells, confirming previous reported findings 6. However, pre-incubation of infected MDMs with wortmannin for 1 hr prior to exposure to ATP resulted in marked inhibition of this bactericidal effect with intracellular BCG viability at 2 hrs (76.8+/- 16.1% p < 0.05), 4 hrs (79.7 ± 5.6% p < 0.05) and 6 hrs (65.5 +/- 4.1%, p < 0.005) remaining significantly higher than in equivalent ATP-treated cells at each time point (Figure 4). To confirm that ATP was acting through P2X₇ to induce cell autophagy, and subsequent bacterial death, infected MDMs were pre-treated with the P2X₇ receptor antagonist, oATP, for 2 hours prior to exposure to ATP. Pre-treatment of cells with oATP was shown to result in almost the complete inhibition of ATP-induced mycobactericidal activity with intracellular BCG viability at both 4 hrs (97.4 ± 3% p < 0.005) and 6 hrs (92.9+/- 1.3% p > 0.005) post-treatment remaining similar to that obtained in non-treated cells, confirming that ATP acts via P2X₇(Figure 4). In total the results suggest that ATP-induced mycobactericidal activity involves P2X₇ activation and the subsequent induction of cell autophagy.

ATP, bzATP, oATP, UTP, wortmannin, ethylene glycol-bis(2-aminoethylether)-N,N,N',N'-tetraacetic acid (EGTA) and calcium chloride (CaCl₂) were purchased from Sigma (Poole, UK). The rabbit anti-LC3 polyclonal antibody and anti-rabbit secondary antibody were obtained from Novus Biologicals and Pierce Biotechnology respectively. The anti-P2X₇ antibody was a kind gift from Dr Gary Buell (Geneva Biomedical Research Institute). Lysotracker Red was purchased from Cambrex Biosciences.

Peripheral blood mononuclear cells (PBMC) were isolated from EDTA blood taken from healthy individuals by standard Ficoll-Hypaque (GE Healthcare Biosciences) gradient separation. Monocytes were then isolated from the PBMC by overnight adherence to 75 ml plastic tissue culture flasks (Corning), washed thrice to remove any non-adherent cells and resuspended by chilling on ice for 30 mins and pipetting with pre-chilled (4°C) PBS. The monocytes were washed and re-suspended (2.5 × 10⁶ cells/ml) in RPMI 1640 medium (Life Technologies, Paisley, U.K.) containing 5% pooled AB⁺ male human serum (First Link, West Midlands, U.K.), 2 mM L-glutamine (Life Technologies) (complete medium) and rh-GM-CSF (90 IU/ml) (Berlex USA).

THP-1 cells were obtained from the American Type Culture Collection (Manassas, VA) and cultured in RPMI 1640 medium + 10% fetal calf serum + 2 mM L-glutamine. The cells were used in a non-activated/differentiated state to maintain their early monocytic phenotype.

For experiments using calcium-free media, the cells were grown in standard Hank's Balanced Salt Solution (HBSS) or in calcium free HBSS + 2 mM EGTA. The latter was added to chelate any residual Ca²⁺.

Stock cultures of M bovis-BCG (Evans strain) were maintained in log phase growth in Middlebrook 7H9 broth (Difco, Detroit, MI) supplemented with 10% Middlebrook ADC enrichment media (Difco) and 0.02% Tween (Difco) at 37°C. The concentration of the BCG stock was determined by direct counting using a Thoma counting chamber (Weber Scientific UK) under dark ground illumination.

GFP-BCG was obtained as a kind gift from Prof D. Young (Imperial College of London, London, U.K). The BCG contained the gene encoding a FACS-optimized GFP protein 23 constitutively expressed under the control of the mycobacterial heat shock protein 60 promoter in a pSMT3 shuttle vector construct 24. Stock aliquots, stored in glycerol at -70°C, were grown to log phase in 7H9 broth supplemented with 10% ADC enrichment medium, 0.2% Tween 80, and 50 μg/ml hygromycin (Sigma, St. Louis, MO).

Following exposure for 30 min to 3 mM ATP, THP1 cells and blood-derived monocytes were washed once in PBS and resuspended in RIPA buffer + protease inhibitor (Sigma, St. Louis, MO) and the cell lysate stored at -80°C prior to assay. SDS-PAGE was performed using a 15% discontinuous gel at 120 V. Following electrophoresis the resolved proteins were transferred onto a PVDF membrane at 80 V for 60 minutes, using a Biorad transfer apparatus. The blot was removed from the transfer apparatus and blocked overnight at 4°C in TBS-T, containing 5% nonfat dried milk. The blot was washed thrice in TBS-T following overnight incubation and then probed with a rabbit polyclonal anti-LC3 antibody (diluted 1:1000 in TBS-T/milk) for 60 minutes. After 3 rinses in TBS-T, peroxidase-conjugated anti-rabbit secondary antibody was applied for 1 hour and the excess antibody removed by washing thrice in TBS-T and the reaction developed using Supersignal West Pico ECL reagents.

In experiments incorporating purinergic receptor agonist and antagonists, oATP (0.3 mM) or anti-P2X₇ antibody (3 μg/ml) were added to the cells for 2 hr/37°C and 1 hr/37°C respectively, prior to addition of ATP (3 mM) for 30 minutes/37°C. UTP (3 mM), and bzATP (3 mM) were similarly added to cells for 30 minutes/37°C before processing them for western blot analysis for LC3 protein expression.

MDMs were grown to sub-confluence on poly-L-lysine coated glass-bottom dishes (MatTek) at approximately 5 × 10⁵ cells/dish for 5 days. GFP-BCG were then added to the macrophages at an MOI of 5:1 and incubated overnight. Following removal of extracellular bacteria by multiple washing, LysoTracker red (75 nM) was added to the cells for 1 hour at 37°C to stain lysosomes. Wortmannin (100 nM) and anti-P2X7 antibody (3 μg/ml) was then added for 1 hr at 37°C before addition of ATP (3 mM) in appropriate wells. Following exposure to ATP for 30 minutes, the cells were washed in RPMI. Cells were imaged live using a Zeiss LSM510 confocal microscope using excitation wavelengths of 488 and 546 nm to visualize GFP and Lysotracker Red respectively. For LC3 staining, MDMs were grown on poly-L-lysine coated glass coverslips for 5 days. Following treatment with ATP, cells were fixed with methanol at -20°C for 30 minutes. Cells were then washed with PBS and incubated in PBS containing 5% donkey serum to block non-specific binding. Cells were incubated with anti- LC3 antibody (1:500) for 1 hour, followed by three washes in PBS and treatment with anti-rabbit secondary antibody, conjugated with AlexaFluor 555 (Invitrogen), for 1 hour. Cells were washed 5 times in PBS and mounted onto slides with Vectashield (Vector Laboratories, CA) as a mounting medium. Fluorescence was visualized by confocal microscopy.

Cells were prepared for transmission electron microscopy (TEM) by pelleting the various cell preparations in 1.5 ml eppendorfs following centrifugation for 1 min/6000 rpm in an eppendorf (Hettich) centrifuge. The cell pellets were then fixed in freshly prepared gluteraldhyde fixative and stored at 4°C until ready for processing. TEM processing of the cell samples was performed at the EM microscopy suite, University of Birmingham according to standard fixation protocols.

Macrophages were grown to sub-confluence (approx 5 × 10⁴cells/well) in 96-well round-bottom microtiter plates (Corning) at 200 μl/well. A total of 2.5 × 10⁵ BCG per well was added (MOI 5:1) and incubated at 37°C/5%CO₂ overnight. Excess BCG was removed by washing thrice in RPMI. Where appropriate, wortmannin (100 nM) and oATP (0.3 mM) were added 1 and 2 hr respectively prior to addition of ATP (3 mM). Intracellular BCG viability following exposure of infected cells to ATP for 30 min, was assessed by ³H-uridine incorporation assay as this provided a much more rapid assessment (within 72 hrs) of the mycobacterstatic effect of ATP as compared to colony forming unit (CFU) assay assessment which takes > 14 days to form countable colonies. The BCG-infected cells were washed and cultured in RPMI +5%AB⁺ serum and at appropriate time points, a total of 150 μl cell supernatant was removed and transferred to a separate 96 well plate. Cell lysis was performed on the remaining cell pellets by addition of 50 ul of a 0.2% saponin solution for 30 min/37°C. After 30 min, 100 μl of Middlebrook 7H9 (+10% ADC supplement) was added to each well of both the cell supernatant and pellet lysates followed by addition of 20 μl/well of a 100 μCi/ml stock of ³H-uridine (Amersham UK). The wells were harvested after 72 hours on a '1205 betaplate scintillation counter' (Wallac) and the combined counts/min of the matched cell supernatant and pellet wells were taken as an indicator of relative bacterial viability.

The Student's t-test was used for all statistical comparisons.