Mice Lacking the Purinergic Receptor P2X5 Exhibit Defective Inflammasome Activation and Early Susceptibility to Listeria monocytogenes

This article is featured in In This Issue, p.553

Listeria monocytogenes is a Gram-positive intracellular bacterium that can cause serious infections in immunocompromised individuals (1). Subsequent to phagocytosis, L. monocytogenes escapes from the phagosomal compartment and gains access to the cytosol, where it grows rapidly within host cells (2). Therefore, the cytosolic innate immune system, including the inflammasome, is critical for host protection. Assembly of the inflammasome, a cytosolic complex of multiproteins, leads to cleavage and activation of caspase-1, which promotes maturation of the proinflammatory cytokines IL-1β and IL-18. Inflammasome formation is triggered by diverse stimuli that are encountered during infection, tissue damage, or metabolic imbalances. L. monocytogenes has been reported to activate the inflammasome, resulting in activation of caspase-1 (3–5). Caspase-1^−/− mice exhibit early susceptibility to L. monocytogenes, characterized by reduced IL-18 release and severely defective IFN-γ production, suggesting the importance of caspase-1–mediated innate immunity in responses against L. monocytogenes (6–9). Cytosolic L. monocytogenes activates the NLRP3 inflammasome (4), and L. monocytogenes total RNA can also stimulate the NLRP3 inflammasome through a variety of mechanisms (3). It has further been found that the Nlrc4 inflammasome contributes to IL-1β production in L. monocytogenes–infected mouse cells (10). In another study, Sauer et al. (11) demonstrated that bacteriolysis within the cytosol resulted in activation of the AIM2 inflammasome, implicating cytosolic survival and avoidance of cell autonomous defenses as important mechanisms of avoiding detection by the AIM2 inflammasome. NLRP6 inflammasome also can be activated by LTA, a molecule produced by Gram-positive bacteria, leading to recruitment of proinflammatory caspases after L. monocytogenes infection (12). Taken together, L. monocytogenes has been demonstrated to engage various inflammasomes, and there are indications that L. monocytogenes–triggered inflammasome activity is critical for immunity. However, it remains unclear by what upstream mechanism L. monocytogenes–triggered inflammasome activation occurs in the context of protective immunity.

P2X receptors are ligand-gated, cation-selective channels with differential permeability to Na⁺, K⁺, and Ca²⁺. Of the seven known P2X receptors (P2X1–7), biological functions have been identified for P2X1, P2X2, P2X3, P2X4, and P2X7 (13). The P2X receptor best characterized for its roles in inflammation and immunity is P2X7, which is highly expressed by virtually all immune cells and is activated by extracellular ATP (14, 15). The discovery of the NALP1 (NLRP1) inflammasome and the identification of additional members of the family (namely NLRP3) enabled placement of P2X7 in a pathophysiological context, providing the molecular mechanism that couples P2X receptor activation to IL-1β processing (16, 17). As of now, P2X7 is understood as one of the most potent activators of NLRP3 inflammasome, capable of initiating caspase-1–mediated processing and release of the proinflammatory cytokines IL-1β and IL-18.

Much less is known about P2X5 in the immune system, with the exception of reported upregulation of P2X5 in CD34⁺ leukemic myeloid cell subpopulations (18) and in activated human T lymphocytes (19). We recently demonstrated a critical role for P2X5 in osteoclast maturation, with P2X5 deficiency resulting in protection against LPS-induced inflammatory bone loss in a manner that involved defects in both inflammasome activation and IL-1β production (20). Therefore, we speculated that P2X5 may also play a role in protective immune responses that are associated with inflammasome activity, such as the immune response to L. monocytogenes infection. As we describe below, we found that P2X5 is a key factor in mounting proper innate immune responses against L. monocytogenes infection in vivo and that P2X5 is required specifically in bone marrow–derived macrophages (BMMs) for L. monocytogenes–induced inflammasome activation.

P2X5-deficient (P2X5^−/−) mice were generated on a C57BL/6J background using P2rx5^−/− sperm obtained from the International Mouse Strain Resources. P2X7-deficient (P2X7^−/−) mice were purchased from The Jackson Laboratory. P2X5^−/−P2X7^−/− mice were generated by crossing P2X5^−/− and P2X7^−/− mice. All mice were maintained and used in accordance with guidelines approved by the Institutional Animal Care and Use Committee at the University of Pennsylvania.

The 10403S wild-type (WT) L. monocytogenes were grown overnight, reinoculated until reaching an OD at 600 nm of 0.3–5 in brain heart infusion medium, and then washed in cold PBS. Mice were infected i.v. with 10⁵ CFU of L. monocytogenes in PBS. In some experiments, mice were treated i.p. with 20 U of apyrase (Millipore Sigma) the day before infection, 20 min before infection, and 6 h postinfection. Blood was collected and serum was isolated for measuring lactate dehydrogenase (LDH) activity using a CytoTox 96 assay kit (Promega) according to the manufacturer’s protocol. Mice were sacrificed 2 and 4 d postinfection. Bacterial loads were determined by plating dilutions of tissue homogenates. For histology, livers and spleens were fixed with 4% paraformaldehyde. Fixed samples were embedded in paraffin, and sections were stained with H&E for microscopic analysis.

IL-12p40, IFN-γ, CCL2, IL-1β, and IL-18 protein levels were measured by ELISA (eBioscience) according to the manufacturer’s protocols. Assays were performed in triplicate for each independent experiment.

Splenocytes from mice infected for 0, 1, and 2 d were cultured with or without 10⁸ bacteria/ml of heat-killed L. monocytogenes stimulation for 4 h in the presence of brefeldin A; then surfaced stained for CD4, CD8a, and CD11b; and finally fixed, permeabilized (Cytofix/Cytoperm; BD Pharmingen), and stained for intracellular TNF-α. Relevant cells were identified after gating from non-CD4, non-CD8, and non-CD11b^int TNF-α^hi lymphocytes/monocytes.

Bone marrow cells were isolated by flushing femurs of mice with RPMI-1640 containing 2% FBS, and spleens were homogenized through 40-μM nylon mesh. Resulting cell suspensions were pelleted by centrifugation at 300 × g. RBCs were lysed and washed three times and counted. Absolute cell numbers were calculated based on the percentage of monocytes from the total cell population acquired by flow cytometry. Single-cell suspensions were blocked with CD16/CD32 (2.4G2) and then stained variously with CD4 (RM4-5), CD8a (53-6.7), B220 (RA3-6B2), NK1.1 (PK136), CD11b (M1/70), Ly6C (HK1.4), Ly6G (1A8), and Siglec-F (E50-2440). Live/Dead fixable dead cell stain kit (Invitrogen) was used to remove dead cells. Monocytes were gated as Ly6C^hiCD11b⁺ from non-CD4, -CD8, -B220, -NK1.1, -Ly6G, and -Siglec-F. All samples were acquired using a FACS LSRII (BD Biosciences) and analyzed using FlowJo software (Tree Star). All Abs were purchased from BD Biosciences except anti-CD4 and anti-Ly6C (eBioscience).

Ly6C^hi CD11b⁺ monocytes sorted on a FACS Aria II (BD Biosciences) were added to the upper chambers of 5-μm transwells (Corning), and 50 ng/ml of CCL2 (Peprotech) was added to the lower chambers. The number of cells that migrated to the lower chamber was enumerated after 2 h.

BMMs were cultured for 7 d in α-MEM supplemented with 10% FBS and recombinant M-CSF (30 ng/ml). A total of 10⁵ BMMs were plated in a 24-well plate 12–16 h before stimulation. The 10403S WT L. monocytogenes, ΔactA mutant, and listeriolysin O (LLO) deleted (Δhly) mutant strains were grown to midlog phase, washed with PBS, and added at a ratio of 10:1 cell (multiplicity of infection [MOI]: 10). Nigericin (20 μM), IL-1β (100 ng/ml; Invivogen), or IL-18 (100 ng/ml; InvivoGen) was added 1 h before infection. One hundred micromolars of ATPγS, 10 U/ml apyrase, 100 μM ADPβS, 100 μM adenosine, and 10 U/ml adenosine deaminase were added 30 min before infection. All were purchased from Millipore Sigma. To measure phagocytosis, cells were lysed with 0.1% Triton X-100 for enumerating initial uptake for 30 min. Two hours postinfection, gentamicin was added to kill the extracellular bacteria, and then 2 h later, cells were lysed with 0.1% Triton X-100, and bacterial CFUs were enumerated by serial dilution.

BMMs were seeded at a density of 10⁶ cells per well in six-well plates and infected with bacteria. The cells were lysed in ice-cold radioimmunoprecipitation lysis buffer (Thermo Fisher Scientific) with protease and phosphatase inhibitor mixture (Roche). Cell lysates were centrifuged to remove debris and quantified by Bradford assay. Equal amounts of lysates (2–50 μg of protein) were fractionated by SDS-PAGE and transferred onto a polyvinyldifluoride (PVDF) membrane. The cleaved forms of caspase-1 and IL-1β were detected using anti–caspase-1 (Adipogen) and anti–IL-1β (AB-401-NA; R&D Systems), with anti–β-actin (Sigma-Aldrich) used for loading control.

Statistical analyses were performed using GraphPad Prism 6 program (GraphPad Software). Differences in mouse survival were assessed using the log-rank (Mantel-Cox) test. Statistical significance was defined as *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

To examine the role of P2X5 in regulating host response to L. monocytogenes, we infected WT and P2X5^−/− mice by i.v. injection. During the early phase of infection (between days 3 and 6), P2X5^−/− mice exhibited significant mortality, whereas all WT mice survived (Fig. 1A). The bacterial burdens in spleens and livers were determined at days 2 and 4, and P2X5^−/− mice had 1–3 log–fold higher titers than WT mice in all tissues (Fig. 1B), confirming that P2X5 is required for bacterial clearance. Histopathological analyses revealed marked lymphocytolysis in P2X5^−/− spleens and more extensive coagulated necrosis with multifocal random microabscesses in P2X5^−/− livers with increased LDH activity (Fig. 1C, 1D). These data indicate that P2X5 is required for host resistance to L. monocytogenes infection.

Because P2X5-deficient mice exhibited early susceptibility to L. monocytogenes infection, we began to investigate whether any critical elements of the anti–L. monocytogenes immune response (2) require P2X5 for expression or function. In particular, early resistance to infection is attributed to the critical roles of MyD88-dependent proinflammatory cytokine release (8), including TNF-α and IL-12, which promote production of IFN-γ (21). Studies with knockout mice show that both IFN-γ (22) and TNF-α (23) are essential for host defense against L. monocytogenes. Therefore, we first measured the serum levels of proinflammatory cytokines upon L. monocytogenes infection. Peak serum levels of IL-12 and IFN-γ appeared to be near normal in P2X5^−/− mice, with IL-12p40 even slightly elevated (Fig. 2A). Because TNF/iNOS-producing–DCs have been reported as the predominant source of TNF-α during L. monocytogenes infection (24), we measured TNF-α production by CD11b⁺ cells harvested from infected spleens and found no significant defect in numbers of TNF-α–producing cells in P2X5^−/− mice (Fig. 2B). Taken together, critical anti–L. monocytogenes proinflammatory cytokines appear to be produced normally in the absence of P2X5, suggesting an alternative explanation for severe susceptibility to L. monocytogenes.

The importance of monocyte recruitment in immune defense against L. monocytogenes infection is well established (25), as shown by experiments that ablate signaling through CCR2 (26, 27). Therefore, we investigated inflammatory monocyte recruitment during L. monocytogenes infection and detected slightly higher peak serum levels of the CCR2 ligand CCL2/MCP-1 in P2X5^−/− mice (Fig. 3A). We examined CCR2-mediated recruitment directly but found no differences in the numbers of WT versus P2X5^−/− sorted Ly6C^hiCD11b⁺ monocytes that migrated in vitro in response to provision of rCCL2 (Fig. 3B). Finally, we infected WT and P2X5^−/− mice with L. monocytogenes and examined recruitment of Ly6C^hiCD11b⁺ inflammatory monocytes as part of the initial infiltrate from bone marrow to splenic white pulp. However, there were no detectable differences in monocyte migration into the tissues between WT and P2X5^−/− mice (Fig. 3C), suggesting that P2X5 does not regulate resistance to L. monocytogenes via monocyte migration.

Roles for P2X receptors in macrophage function during pathophysiological conditions have previously been reported. P2X7 deficiency results in defective Chlamydia trachomatis (28) killing in macrophages, and P2X4 deficiency causes higher susceptibility in response to Escherichia coli that is associated with uncontrolled bacterial killing in macrophages in vitro (29). The role of P2X receptors may depend on the type of bacterium and cell type studied. To specifically evaluate the relationship between L. monocytogenes killing capacity by macrophages and P2X5 expression, we differentiated BMMs from WT and P2X5^−/− mice and then performed in vitro infection with L. monocytogenes, followed by quantification of remaining live bacteria that were taken up by BMMs. Killing capacity was found to be significantly reduced in P2X5^−/− versus WT BMMs (Fig. 4). Decreased killing was not due to differences in either initial phagocytosis or intracellular growth, as similar numbers of colonies were counted within WT and P2X5^−/− BMMs after 30 min incubation and growth curves for bacteria within WT and P2X5^−/− BMMs remained similar throughout the entire course of infection (Supplemental Fig. 1). Furthermore, levels of cell death exhibited by WT and P2X5^−/− BMMs during L. monocytogenes killing assays were similar (data not shown). To attempt to identify specific mechanisms of P2X5-dependent L. monocytogenes killing, we also tested killing capacity using several L. monocytogenes mutant strains, including LLO and actA mutants, and found that whereas the actA mutant resulted in a similar killing defect as WT L. monocytogenes, LLO mutant killing was similarly reduced in both WT and P2X5^−/− BMMs (Fig. 4). Upon entry into the BMMs, L. monocytogenes is initially contained within a phagocytic vacuole, and LLO is necessary for allowing L. monocytogenes to escape into the cytosol (30). Once in the cytosol, L. monocytogenes become motile by using actin tails generated by ActA, which allows L. monocytogenes to spread to neighboring cells (31). Therefore, these mutants allow us to infer that the killing defect in the vacuole is P2X5 independent, whereas cytosolic killing is P2X5 dependent. In addition, because LLO is believed to be the primary L. monocytogenes–associated molecule that triggers the inflammasome (32, 33), similarly defective LLO mutant killing in both WT and P2X5^−/− BMMs suggests that P2X5 may be required for inflammasome-associated killing of L. monocytogenes.

Because the P2X5-dependent defect in L. monocytogenes killing implicated the inflammasome, we next sought to determine whether P2X5 is required for L. monocytogenes–induced inflammasome activation. In vitro infection of WT BMMs with L. monocytogenes resulted in caspase-1 activation, as represented by cleavage of procaspase-1 to caspase-1p20, but P2X5^−/− BMMs exhibited significant decreases in both caspase-1 activation and maturation of pro–IL-1β to the mature IL-1βp17 form (Fig. 5A). These defects were correlated with IL-1β protein levels measured in culture supernatants by ELISA (Fig. 5B) as well as with another inflammasome-dependent cytokine, IL-18, which also exhibited lower production in P2X5^−/− BMMs (Fig. 5B). These results strongly suggest that P2X5 is required for L. monocytogenes–triggered inflammasome activation in BMMs. To investigate whether P2X5-dependent L. monocytogenes–triggered inflammasome activation is related to defective L. monocytogenes killing in P2X5^−/− BMMs, we treated WT and P2X5^−/− BMMs with the K⁺ ionophore nigericin, a direct chemical activator of the inflammasome (34), in the context of an in vitro L. monocytogenes killing assay. Nigericin-mediated rescue of L. monocytogenes killing in P2X5^−/− BMMs suggests both that inflammasome activation downstream of P2X5 remains intact and that P2X5^−/− BMMs are capable of normal L. monocytogenes killing if P2X5-dependent activation of the inflammasome is chemically bypassed (Fig. 5C). Finally, we showed that L. monocytogenes killing by P2X5^−/− BMMs was partially rescued by supplementing cultures with the inflammasome-dependent cytokines IL-1β or IL-18 for 1 h prior to L. monocytogenes infection (Fig. 5D). Together, these data suggest that BMM-expressed P2X5 is necessary for optimal inflammasome activation and production of the inflammasome-dependent cytokines IL-1β and IL-18 and that P2X5-dependent L. monocytogenes killing requires inflammasome activation and production of IL-1β or IL-18.

P2X7, which is highly expressed on various immune cells, including BMMs, is the P2X family member most associated with immune regulation and has been implicated in responses to bacterial infections (35, 36, 37). It is therefore reasonable to investigate whether P2X5-dependent anti–L. monocytogenes immunity is linked to P2X7 function. To examine the role of P2X7 in regulating host response to L. monocytogenes and how P2X7 expression might affect P2X5-dependent anti–L. monocytogenes immunity, we infected WT, P2X7^−/−, and P2X5^−/−P2X7^−/− mice by i.v. injection. Whereas P2X5^−/− and P2X5^−/−P2X7^−/− mice exhibited significant mortality (between days 3 and 5), all WT and P2X7^−/− mice survived (Fig. 6A), suggesting that P2X7 is dispensable for early systemic immunity to L. monocytogenes. Investigating in vitro killing of L. monocytogenes, we found that unlike P2X5^−/− BMMs, L. monocytogenes killing capacity by P2X7^−/− BMMs was normal (Fig. 6B). These results led us to investigate whether control of L. monocytogenes requires the presence of the established P2X/P2X7 purinergic receptor agonist, extracellular ATP. To do so, we treated WT and P2X5^−/− mice before and after L. monocytogenes infection with the ATP-hydrolyzing agent apyrase and found that apyrase treatment caused no increase in mortality of WT mice (Fig. 6C). More surprisingly, apyrase treatment resulted in complete rescue of L. monocytogenes–infected P2X5^−/− mice (Fig. 6C), suggesting that ATP-mediated signaling is not required for early anti–L. monocytogenes immunity but even possibly that apyrase-induced production of ATP metabolites can compensate for loss of P2X5-dependent anti–L. monocytogenes function. To investigate the role of extracellular ATP and ATP metabolites on L. monocytogenes killing by BMMs, we performed in vitro killing assays using WT and P2X5^−/− BMMs in the presence of apyrase, extracellular ATP, and ATP metabolites and found, consistent with our in vivo infection findings, that apyrase rescues defective L. monocytogenes killing by P2X5^−/− BMMs (Fig. 6D). By contrast, treatment with a nonhydrolyzable form of ATP (ATPγS) had no effect on either WT or P2X5^−/− killing or any significant effect on apyrase-mediated rescue (Fig. 6D). Consistent with our speculation, however, addition of ATP metabolites ADPβS or adenosine combined with deaminase did induce partial rescue of defective P2X5^−/− killing (Fig. 6D). Together, these results suggest that P2X5 does not use conventional ATP-triggered signaling to effect anti–L. monocytogenes immunity.

In our previous studies, we demonstrated a relationship between P2X5 and inflammation by showing the regulation of inflammasome activity in inflammatory bone loss (20, 38). In this study, we sought to identify the activation pathway in L. monocytogenes–induced inflammatory conditions. We clearly showed the role of P2X5 in controlling L. monocytogenes infection in vivo and elucidated a mechanism for control of intracellular L. monocytogenes bacteria in BMMs in vitro, whereby P2X5 specifically increases L. monocytogenes killing but not phagocytosis. In addition, we provide evidence that BMM-expressed P2X5 functions to activate caspase-1 and IL-1β and IL-18 maturation in response to L. monocytogenes in association with optimal bacterial clearance. Of note, the levels of IL-1β and IL-18 in the serum of L. monocytogenes–infected P2X5^−/− mice were found not to be reduced compared with infected control mice (Supplemental Fig. 2). Therefore, consistent with previous reports that neither IL-1β nor IL-18 is required for early resistance to L. monocytogenes (39, 40), our findings show that rescue of defective L. monocytogenes killing by P2X5^−/− BMMs via addition of exogenous IL-1β or IL-18 may be indicative of a local rather than systemic defect. It is also possible that increased systemic IL-18 production in L. monocytogenes–infected P2X5^−/− mice might be due to inflammatory factors released because of uncontrolled L. monocytogenes infection.

Caspase-1 activation is controlled by the inflammasome, a multiprotein complex formed by NLR proteins and the adaptor ASC (41). NLRP3, a critical inflammasome component, is activated by microbial stimuli such as L. monocytogenes. One of the favored models of NLRP3 inflammasome activation is that P2X7-dependent pore formation by pannexin-1 allows extracellular NLRP3 agonists to enter the cytosol and directly engage NLRP3 (42). However, it has also been reported that P2X7 receptor is differentially required for caspase-1 activation induced by intracellular and extracellular bacteria (43). Activation of inflammasomes by intracellular bacteria, including L. monocytogenes, proceeds normally in the absence of P2X7 receptor-mediated cytoplasmic K⁺ perturbations. Consistent with this model, we found that unlike P2X5^−/− mice, P2X7^−/− mice are resistant to L. monocytogenes infection and, further, that P2X7^−/− BMMs exhibit normal L. monocytogenes killing. Because P2X7 is highly expressed on BMMs, as well as other immune cells, the fact that it is dispensable in the same context where P2X5 is required suggests distinct functional activity that is unique to each receptor. Interestingly, we found that early immunity to L. monocytogenes infection and BMM killing of L. monocytogenes did not require the presence of the established P2X/P2X7 agonist extracellular ATP. Although ATP metabolites may be capable of overcoming P2X5-associated defects, it remains unclear what L. monocytogenes–associated factors are necessary for triggering P2X5-dependent immunity. Regardless, our findings suggest that P2X5 is required to induce L. monocytogenes–triggered inflammasome activation in BMMs and is the first example, to our knowledge, of an upstream activating receptor of caspase-1 that is specifically required for anti–L. monocytogenes immunity.

Numerous factors identified as being critically important to early anti–L. monocytogenes immunity, including MyD88 (7, 8), caspase-1 (6), and IFN-γR (21, 22), are associated with production and/or function of IFN-γ. In addition, monocyte recruitment, which is linked to CCR2 expression, is a critical function in early anti–L. monocytogenes immunity (25). However, although P2X5^−/− mice exhibit severe susceptibility to L. monocytogenes, they produce normal levels of IFN-γ and exhibit normal monocyte recruitment. These observations are notable in comparison with the phenotype of caspase-1^−/− mice, which show somewhat similar susceptibility to L. monocytogenes as P2X5^−/− mice but also exhibit a severe defect in serum IFN-γ production (6). These observations suggest first that there may be other L. monocytogenes–triggered receptors that activate caspase-1 in vivo, possibly expressed on cells other than macrophages, that are responsible for other functions (e.g., IFN-γ and IL-18 production) in the context of L. monocytogenes infection. Second, given that P2X5^−/− mice show early susceptibility to L. monocytogenes, even though they exhibit apparently normal IFN-γ production and monocyte recruitment, it suggests that IFN-γ production and monocyte recruitment together are not sufficient for early resistance to L. monocytogenes and that P2X5-dependent responses may highlight a novel requisite component of early anti–L. monocytogenes immunity. Future studies focused on cell-specific expression of P2X5 will be important for better understanding how this little characterized purinergic receptor plays a critical role in anti–L. monocytogenes immunity. Finally, we believe the findings we have presented in this article are important for elucidating not only how P2X receptors regulate immune responses but also how antibacterial immunity relies on various specialized mechanisms of purinergic signaling under different pathophysiologic conditions.

We thank Dr. Sunny Shin, Dr. Hyunsoo Kim, and Dr. Noriko Takegahara at the University of Pennsylvania for helpful discussion and critical reading of the manuscript.

The authors have no financial conflicts of interest.