Introduction

Intracellular pathogens invade, survive and replicate in mammalian cells, modulating host-cell membrane trafficking and cytoskeletal dynamics in order to establish persistent infection in the mammalian host [1–3]. Macrophages are a frequent target of microbial infections, and they respond to microbial invasion by producing factors such as reactive nitrogen and oxygen intermediates (RNIs and ROIs) that have strong microbicidal activity [4]. But many pathogens have evolved different strategies for avoiding destruction by the macrophage. Some intracellular pathogens, such as Mycobacteria, inhabit a compartment whose endocytic maturation is delayed, while Chlamydiae survive within a membrane-bound vacuole that avoids fusion with lysosomes and maintains a neutral pH [5, 6]. Unlike Mycobacteria and Chlamydia, the protozoan parasite Leishmania thrives in a parasitophorous vacuole that has an acidic pH and high hydrolytic activity, and Trypanosoma cruzi actively induces its uptake into lysosome-like host-cell vacuoles, from which, the parasite rapidly escapes into the cytosol [7–10].

At the end of their infection cycle, each of these intracellular parasites is released from the host cell, triggering macrophage death and inducing local inflammation, accompanied by the possible release of ATP, as shown for macrophages infected with Mycobacterium tuberculosis [11]. As extracellular ATP (ATPe) can be used by neighbouring macrophages as ammunition to inhibit infection (described below), many intracellular pathogens such as Mycobacterium bovis BCG, Vibrio cholerae, M. tuberculosis, T. cruzi and Leishmania also secrete or express on their outer surface enzymes that degrade or synthesise nucleotides [12–15]; and microbial enzymes that consume or produce ATP are considered as virulence factors for M. tuberculosis, Leishmania amazonensis and T. cruzi [14, 16–18].

Macrophages activate microbicidal pathways and contribute to inflammation after the ligation of purinergic P2 receptors by ATPe. This review will, therefore, describe some of the P2-dependent mechanisms used by macrophages to eliminate infection by intracellular bacteria and protozoan parasites, and, whenever possible, will correlate these findings with host susceptibility and resistance to infection.

Effect of ATPe on macrophage infection by M. tuberculosis

The first evidence for an involvement of ATPe in the control of intracellular infections came from the laboratory of Kaplan, who demonstrated in 1994, that ATPe-mediated apoptosis in Mycobacterium tuberculosis-infected macrophages is associated with the inhibition of mycobacterial infection [19]. This early study also showed that ATPe-induced macrophage apoptosis, but not H2O2-induced necrosis, is associated with the killing of the intracellular mycobacteria. These findings were confirmed by Lammas et al., who proposed that the ATPe-induced elimination of BCG-infected human macrophages are mediated by the purinergic receptor, P2X7, through a mechanism independent of both RNIs and ROIs [20]. In comparison with other ligands that can trigger the lysis of macrophages, including complement-mediated cytolysis, Fas ligation and CD69 activation, only ATPe treatment could stimulate the death of both host macrophages and intracellular mycobacteria. Subsequently, several laboratories confirmed that the P2X7 receptor plays a role in limiting infection in murine and bovine macrophages infected with mycobacteria, and in human macrophages infected with BCG bacillus and virulent strains of M. tuberculosis [11, 21–23].

In addition to the P2X7 receptor, other P2 receptor subtypes, possibly P2Y, are apparently involved in ATPe-mediated bactericidal activity in macrophages [11, 21, 23]. Experiments using macrophages derived from P2X7-deficient mice revealed that ATPe stimulates the production of reactive species such as RNIs equally well in both wildtype and P2X7-deficient macrophages [11]. Moreover, it was found that ATPe induces bactericidal effects in the macrophages better than BzATP (the most potent known agonist for the P2X7 receptor), suggesting that P2X7 receptors are necessary, but not sufficient, for maximal ATPe-dependent killing of intracellular M. tuberculosis by human and bovine macrophages [21, 22]. Lammas et al. have further observed that the ATPe activity is potentiated by extracellular Zn2+ [20]. This effect was initially ascribed to the P2X7 receptor, but now, P2X7 activity is known to be blocked by extracellular Zn2+, while the activity of another purinergic receptor, P2X4, is potentiated by Zn2+ [24]. Since macrophages express functional P2X4 receptors [25] and inflammatory mediators can upregulate this receptor on macrophages [26], it is likely that both P2X7 and P2X4 are involved in the ATPe-induced killing of M. tuberculosis in macrophages.

More recent reports have also confirmed the predominant role of the P2X7 receptor in mycobacterial clearance, extending these results to show that loss-of-function polymorphisms in human P2X7 receptors lead not only to reduced ATPe-induced apoptosis, but also to impaired ATPe-induced killing of intracellular mycobacteria (BCG) by macrophages [27–29]. Nonetheless, more experiments will be required to elucidate the role that P2X7, possibly in conjunction with other P2 receptors, may play in the killing of intracellular mycobacteria in vivo, since P2X7-deficient mice control lung infection as well as wildtype mice after low-dose aerosol infection with virulent M. tuberculosis [30].

Cellular mechanisms of ATPe-induced mycobacterial killing

ATPe ligation of P2X7 on macrophages results in a variety of different cellular effects, including the activation of phospholipase D (PLD), maturation and release of interleukin-1β (IL-1β), generation of macrophage polykarions, modulation of lipopolysaccharide (LPS) induced macrophage activation through modulation of iNOS expression and NO production, and the induction of macrophage death by necrosis and/or apoptosis [31–40]. The original findings from the Kaplan laboratory suggested that the apoptosis of macrophages is necessary for ATPe-mediated killing of intracellular bacteria [19], but it was later established that the ATPe-induced killing of mycobacteria in human and mice macrophages can occur without macrophage death, through a pathway requiring PLD activation, the acidification of phagosomes and phagosome–lysosome fusion [21] (Fig. 1) [23, 41]. A more recent study showed that, when combined, two loss-of-function polymorphisms in human P2X7 receptors impair ATPe-mediated apoptosis, despite the normal killing of BCG bacillus [28], reinforcing the view that the apoptosis of macrophages is not necessary for the elimination of mycobacteria.

Fig. 1
figure 1

ATP released from infected cells undergoing necrosis or sites of inflammation can bind to the P2X7 receptor on neighbouring macrophages and other cells. Ligation of the P2X7 receptor initiates signalling through several pathways, which result in the activation of caspase-1, activation of phospholipase D (PLD) and, ultimately, cell death. Caspase-1 activation stimulates the inflammatory response by the cleavage of pro-IL-1β and secretion of the mature cytokine. PLD activation modifies membrane trafficking in the cell, which can induce fusion between lysosomes and vacuoles, harbouring intracellular pathogens such as mycobacteria and chlamydiae. The death of macrophages is partly necrotic, which may amplify inflammation even further, through the ligation of P2X7 receptors on other cells. Some extracellular ATP may also bind to other P2 receptors, which may inhibit infection by activating as-yet-uncharacterised pathways

Full size image

In fact, it was recently shown [42] that cyclosporine A, an inhibitor of the mitochondrial permeability transition, increases the survival of human monocyte-derived macrophages infected with M. tuberculosis, restores P2X7 function and enhances antimycobacterial activity. Conversely, M. tuberculosis has developed mechanisms to evade P2X7-triggered mechanisms, since it can secrete a nucleoside diphosphate kinase that produces ATP and kills macrophages through a P2X7-dependent mechanism [14]. New experiments are needed to clarify the role played by ATPe in killing intracellular pathogens and inducing host-cell death.

Inhibition of chlamydial infection in macrophages

The Chlamydia species are obligate intracellular bacteria that infect mainly epithelial mucosa, where they survive within intracellular vacuoles that avoid fusion with host-cell lysosomes [6, 43]. Different strains of C. trachomatis are responsible for the infection of genital and ocular tissue in humans [44–48]. C. pneumoniae is a common cause of community-acquired pneumonia in humans and is associated with an increased risk for atherosclerosis [44, 49]. Both C. trachomatis and C. pneumoniae can invade epithelial cells and macrophages in vitro and in vivo [49, 50].

It was recently shown that ATPe inhibits the infection of macrophages by the murine species of C. trachomatis, C. muridarum, through a mechanism that required PLD activation and fusion between lysosomes and the Chlamydia vacuoles [51]. The effect of ATPe was dependent on the presence of the P2X7 receptor, since there was no PLD activation nor killing of chlamydiae in infected macrophages that had been isolated from P2X7-deficient mice. Although P2X7 ligation also led to macrophage death, the inhibition of PLD prevented chlamydial killing but had no effect on macrophage death, suggesting that PLD activation was directly responsible for the inhibition of chlamydial infection (Table 1). Moreover, fusion between lysosomes and chlamydial vacuoles preceded macrophage death, further strengthening the conclusion that the killing of chlamydiae is independent of host-cell death [51].

Table 1 Effect of extracellular ATP on intracellular pathogens
Full size table

Conversely, as the activation of P2X7-dependent pathways is deleterious for Chlamydia, both directly and through the demise of the macrophage, the intracellular pathogen has also evolved a mechanism for protecting its host cell. Thus, the infection of macrophages with a related species, C. psittaci (also known as C. caviae), inhibits partially ATPe-mediated macrophage death [52]. While the molecular basis for host-cell protection remains to be investigated, chlamydial infection decreases partially the ability of ATPe to induce plasma–membrane permeabilisation and calcium fluxes [52]. Chlamydiae, therefore, resemble other pathogenic bacteria and protozoan parasites that attempt to protect themselves and the host cell by degrading nucleotides or hydrolysing ATP.

Effects of ATPe on leishmaniasis

Leishmaniasis is used to describe several diseases caused by the obligate intracellular protozoan parasite Leishmania, which infects mainly macrophages [53]. The diseases range from self-healing cutaneous lesions to visceral and potentially fatal disseminating infection. Leishmania infections are found in 80 countries, with a prevalence of 12 million human cases. The development of different clinical forms is associated with both the immunological status of the host and the parasite species [53]. The expression of the P2X7 receptor has recently been examined during Leishmania infection, revealing upregulation of the receptor during both in vivo and in vitro infection with L. amazonensis [54]. These changes were correlated with functional responses, as reflected by an increase in ATPe-mediated plasma–membrane permeabilisation and host-cell apoptosis ([54] and Chaves et al. (manuscript in preparation)) (Table 1). The increase in ATPe-induced membrane permeabilisation was also observed in spleen macrophages isolated from mice infected with L. donovani, suggesting that this may be a general phenomenon relevant for all Leishmania infections.

It has been proposed that intracellular infection by Leishmania donovani inhibits macrophage apoptosis induced by growth factor deprivation [55]. In contrast, there is an increase in the ATPe-mediated apoptosis of macrophages infected with Leishmania (our unpublished data), consistent with the increases in ATPe-mediated membrane permeabilisation. Thus, despite the production of ecto-ATPases by Leishmania [16], this strategy is not sufficient to protect the host cell against the infection-dependent upregulation in P2X7 expression.

We have observed that the presence of ATPe during L. amazonensis infection does not interfere with Leishmania invasion, but the ATPe treatment of macrophages that are already infected with L. amazonensi leads to a decrease in Leishmania survival (our unpublished observations). In addition, we observed that ATPe has no effect on the viability of extracellular Leishmania promastigotes. In fact, some nucleotides, such as UTP, stimulate the proliferation of promastigotes. It is also worthwhile noting that the more infective forms of L. amazonensis express more magnesium-dependent ecto-ATPase on their membranes than less virulent Leishmania, leading to the proposal that the ecto-ATPase should be viewed as a virulence factor of the parasite [16].

The cellular pathway allowing ATPe to decrease Leishmania infection remains to be determined, but must be different from the PLD activation observed during Chlamydia and Mycobacterium infection [41, 51], since both Chlamydia and Mycobacterium inhibit phagolysosome formation and acidification [4, 6, 8], while Leishmania survives well in acidic phagolysosomes [56]. Interestingly, Leishmania lipophosphoglycans, which promote parasite survival, act by perturbing MAPKinase signalling in macrophages to inhibit macrophage IL-1β [57]. This might be relevant for the involvement of P2 receptors in the escape mechanisms used by Leishmania, since several P2 receptors are connected to MAPKinase pathways [58].

The involvement of ATPe in Chagas–disease

Chagas–disease is caused by the facultative intracellular protozoan pathogen, T. cruzi. The disease is a chronic inflammatory condition characterised by cardiomyopathy and digestive disorders [59, 60]. T. cruzi infection affects over 17 million people in endemic areas of Latin America, leading to 45 thousand deaths per year [61]. The involvement of ATPe signalling was recently examined during the acute phase of T. cruzi infection. Thymus atrophy occurs during the acute phase of infection but the thymus recovers weight and cellularity during the chronic phase [62]. These alterations do not appear to be associated with stress or glucocorticoid release [63]. It has recently been observed that thymocytes are sensitive to ATPe-induced membrane permeabilisation and host-cell death only during the atrophy phase of infection, with CD4+/CD8+ double-positive thymocytes being the most sensitive subpopulation of thymocytes [64] (Table 1). Since the phenomenon of ATP-induced permeabilisation can be blocked by the P2X7 inhibitors, oxidised ATP and Mg2+, and the P2X7 agonist, BzATP, was more potent than ATP, we proposed that the increased sensitivity to ATPe may be responsible, at least in part, for the thymocyte clearance and thymic atrophy observed during the acute phase of Chagas’s disease [64].

It is, therefore, reasonable to suppose that ATPe plays an important role in the T. cruzi infection cycle. In this context, it is worth noting that the ecto-ATPases produced by these parasites have been associated with strain virulence [17].

P2X7 receptors and inflammation

The stimulation of P2X7 by ATPe leads to caspase-1 activation, cleavage of pro-IL-1β and the secretion of mature IL-1β [66] (Fig. 1). The P2X7 receptor is also upregulated in macrophages by inflammatory cytokines, such as interferon-γ (IFN-γ) and tumour necrosis factor-α (TNF-α), and lipopolysaccharide (LPS). LPS, a surface component of most Gram-negative bacteria, and IFN-γ act synergistically to upregulate P2X7 receptor function and P2X7 mRNA in the human monocyte cell line, THP-1, and human macrophages [34, 37]. In fact, P2X7 receptors contain several motifs that are homologous to motifs from other receptors known to be involved in protein–protein interactions and LPS binding [67]. The upregulation of P2X7 expression in monocytes by TNF-α, LPS and IFN-γ is consistent with the ability of these cytokines to act as inflammatory mediators, and these effects are markedly attenuated by coincubation with prostaglandin E2 (PGE2) or the membrane-permeable cAMP analogue, dibutyril cAMP [68]. It is tempting to speculate that the temporal sequence of macrophage exposure to pro-inflammatory activators and anti-inflammatory stimuli (such as PGE2) might regulate not only receptor expression, but also downstream signalling by the P2X7 receptors.

It is known that macrophages express adenosine receptors, which are expressed during the differentiation of monocytes to macrophages and may influence phagocytosis [69]. Moreover, the treatment of macrophages with IFN-γ upregulates expression of the adenosine receptor, A2B, and the activation of A2B receptors is involved with the deactivation of macrophages, possibly through an increase of cAMP [70]. Therefore, the extracellular nucleotides may be involved with activation and a feedback mechanism for macrophage deactivation, depending on the timing and type of nucleotide released during infection. In this context, \studies with P2X7-deficient mice have reinforced the view that P2X7 receptors are involved in inflammation. Thus, disruption of the P2X7 receptor gene is associated with less severe disease in an arthritis model [71] and studies of chronic inflammation and neuropathic pain [72]. These results led to the hypothesis that the P2X7 receptor, through the regulation of IL-1β production and secretion, plays a common, early role in the development of pain of neuropathic and inflammatory origin [71, 72]. Additionally, a recent study demonstrates that the inhibition of the P2X7 receptor attenuates fever and cytokine responses induced by LPS in rats [73]. Finally, the P2X7 receptor is present and its expression is modulated by inflammation in sites of chronic inflammation [74].

All of these findings reinforce the conclusion that P2X7-dependent signalling plays a significant role in host responses during various types of inflammatory disease. However, the involvement of P2X7 during disease is complicated by the presence of other P2 receptors, which may also contribute to either pro- or anti-inflammatory immune responses. Moreover, ATP can negatively regulate Toll-like receptor signalling, suppress LPS-induced MCP-1 and TNF-α, and augments IL-10 production in human monocytes [75]. Many infectious agents that survive in these immune effector cells may have evolved complex nucleotide-based strategies to evade the immune system. Dissection of these strategies may lead us to a better understanding of the role of nucleotide signalling in the immune response and to the development of new approaches to combat infectious diseases.

Concluding remarks

As large concentrations of ATP are present outside of the cell only when the cell is damaged or is part of an inflamed tissue (Fig. 1), it has been proposed that ATPe may function as a generic ‘danger signal,–which could alert the immune system to the presence of any type of intracellular pathogen that induces host-cell death [76–78]. Different intracellular pathogens, such as bacteria and protozoan parasites, express ecto-ATPases or other mechanisms to either inhibit or enhance ATPe-mediated death of their host cell, suggesting that ATPe may have been used by the host as an ancient danger signal, to which, intracellular pathogens have been exposed since the early evolution of the immune system. In this context, the ability of some intracellular pathogens to inhibit the ATPe-dependent response mediated by P2 receptors could be an example of adaptation of the pathogenic invaders to the immune response, and may help to explain why the most virulent pathogens express high ecto-ATPase levels on their surface [14–17]. Given the availability of animals that are deficient in P2X7 and other P2 receptors, further research in the future will, thus, need to address the relevance of P2-dependent immune mechanisms in controlling infections in whole organisms.