P2 receptors and immunity

Abstract

Immune cells express receptors for extracellular nucleotides named P2 receptors. P2 receptors transduce signals delivered by nucleotides present in the extracellular environment. Accruing evidence shows that purinergic signalling has a profound effect on multiple immune cell responses such as T lymphocyte proliferation, chemotaxis, cytokine release, phagocytosis, Ag presentation and cytotoxicity. This makes P2 receptors an attractive target for the therapy of immuno-mediated disease and cancer.

1. Nucleotides as extracellular messengers

Adenosine triphosphate (ATP) is probably one of the oldest signalling molecules appeared during evolution of living organisms. Albeit it is basically impossible to gather experimental proofs, it would not be surprising that ATP was used as a signalling molecules at the same time as it started to be exploited as an intracellular energy source. There are several reports of the activity of extracellular ATP in primitive organisms such as bacteria, algae and slime moulds [1]. Evidence for a signalling role of extracellular nucleotides is compelling in lower invertebrates and overwhelming in mammals [2]. ATP, and its degradation product adenosine, are well recognized neurotransmitters in both the central and peripheral nervous system, but their signalling role in other systems is not yet fully acknowledged. In this respect, it is ironic that identification of ATP as the high energy intermediate in muscle and first description of extracellular ATP and adenosine effects on heart beat occurred in the same year, 1929 [3,4], but while the role of ATP as an intracellular high energy intermediate was plainly accepted, it took almost a century to acknowledge its participation in extracellular signalling. This is even more surprising from the immunologist standpoint. Students of inflammation and innate immunity have always been busy investigating the elementary signals released by distressed cells that alert the immune system of an impending danger. Nowadays, it is a solid observation that detection of a foreign microorganism is not by itself sufficient to start inflammation, but recognition of a damage is also needed. In fact, recognition of damage is even more important than detection of a foreign agent, as clearly shown by the occurrence of sterile inflammation.

2. ATP as a DAMP

Cells of innate immunity recognize PAMPs (Pathogen Associated Molecular Patterns) released by invading microorganisms, but they also need to detect DAMPs (Damage Associated Molecular Patterns) in order to be fully activated [5,6]. All DAMPs share the unique feature of being virtually absent in the extracellular environment in healthy conditions, while on the contrary are released quickly upon cell damage. Since their extracellular concentration is normally close to nil, even a small leakage of these compounds generates a large signal that can be quickly detected by nearby immune cells. Several intracellular molecules have been so far listed in the DAMP family: high mobility group box 1 (HMGB1) protein, heat-shock proteins, nucleosomes, DNA and ATP [7]. Several features make ATP an ideal DAMP, probably better than the other molecules so far proposed to this role. First of all its huge outward-directed transmembrane gradient (cytosol concentration in the millimolar range, extracellular concentration in the low nanomolar range) makes it very easy to rapidly generate a large extracellular signal following even minor cellular insults; secondly, ATP is highly diffusible through the aqueous extracellular environment thanks to its charges (2 or 4 negative charges depending on the pH and extracellular cation concentration); thirdly, ATP binds with varying affinity to a large family of specific cognate receptors, a feature that confers to ATP-based signalling a unique plasticity; fourthly, the ATP signal can be quickly terminated by ubiquitous ecto-ATPases, thus fulfilling a basic requirement of every bona fide messenger molecule, i.e. rapid inactivation when the message is no more needed [2,8].

In the initial phases of inflammation DAMP signalling is exquisitely needed to activate dendritic cells (DCs) and therefore to modulate the ensuing immune response, thus it is anticipated that an additional requirement of ATP in order to be considered a DAMP is that this nucleotide acts on the DCs to potentiate antigen presentation and direct the evolution of the immune response. Several reports now confirm the key immunoregulatory role of extracellular ATP proposed in early studies [9–12]. As any extracellular, membrane-impermeant, mediator ATP is sensed by plasma membrane receptors named P2 receptors. In turn, adenosine is sensed by the other members of the purinergic receptor family named P1 receptors. P2 receptors are further subdivided into P2Y (G protein coupled) and P2X (intrinsic ion channels). Four P1 receptor subtypes (A1, A2A, A2B and A3), eight P2Y (P2Y1,2,4,6,11,12,13,14) and seven P2X (P2X1-7) subtypes are known [13,14]. Basically all P1 and P2 receptor subtypes are expressed by immune cells, in a cell type- and differentiation-dependent fashion.

As it is often the case in the signalling field the fundamental question of the detection and measurement of the putative mediator (ATP) in the extracellular environment has remained unanswered for a long time due to lack of suitable tools to measure the ATP concentration in the interstitial fluid in vivo.

Measuring ATP in the extracellular environment avoiding at the same time any possible cell damage that might by itself trigger ATP release is a daunting task, something comparable to Eisenberg uncertainty principle, if “parva licet componere magnis”. It is in fact a common observation that even minor cell perturbations, such as for example rinsing cell monolayers with culture medium, trigger large ATP release, thus we can imagine what happens any time investigators stick an electrode into a tissue to measure adenosine or ATP release [15]. This intrinsic difficulty in ATP measurement has been tackled (and hopefully solved) by our group with the engineering of a modified luciferase targeted to and stably expressed on the cell plasma membrane (catalytic site outside) [16]. We anticipated that such a probe, named pmeLUC (plasma membrane luciferase) would allow measurement of ATP in the pericellular milieu. This prediction has been fulfilled in numerous experimental settings [17–19], thus making pmeLUC the probe of choice to measure extracellular ATP in vivo. The convergent indication stemming from the application of pmeLUC to extracellular ATP measurement has been that while concentration of this nucleotide in healthy tissues is negligible, and in any case below detection level, at sites of inflammation or within tumor microenvironment the concentration can be as high as hundreds micromolar [17,18]. This concentration is sufficient to activate even the low affinity P2X7 receptor and to generate large amounts of the anti-inflammatory agent adenosine. The obvious question then arises as to how this large extracellular ATP concentration is generated and exploited by inflammatory cells to orient themselves through the inflamed tissue, reach and destroy the causative agent, and finally repair the tissue.

3. Extracellular ATP: an immunomodulatory factor

For a long time it has been assumed that the only pathway for cellular ATP release was plasma membrane damage or overt cell injury. We now know that all cells are capable of non-lytic ATP release and that this autocrine/paracrine purinergic stimulation has a key role in the pathophysiology of immune cells. Several mechanisms are responsible for ATP release: exocytosis of ATP-containing granules, plasma membrane carriers, large conductance channels such as connexins and pannexins, and P2 receptors themselves such as P2X7 [20,21]. It is not clear which ATP release mechanism predominates in immune cells, but it is likely that multiple release pathways are involved, possibly also depending on the activating stimulus and the given pathophysiological condition. Some ATP release pathways are of particular interest because they might trigger an autoregenerative ATP release loop resulting in spreading and amplification of the ATP signal. This is the case of ATP release through P2X7, as released ATP might feed back onto the P2X7 receptor itself to keep it in an open state and thus allow further ATP release. Release of ATP at the inflammatory site generates an ATP concentration gradient that might be exploited by inflammatory cells for chemotaxis. There is evidence that chemotactic peptides trigger ATP release from the leading edge of neutrophils and that ATP activates P2Y2 receptors that in turn potentiate the chemotactic signal [21]. Chemotaxis implies sensing and moving through the chemotactic gradient. Activation of neutrophils movement through the chemotactic gradient seems to be due to adenosine accumulation at the leading edge of the chemotacting neutrophil and the following A3 receptor stimulation [22]. More recent studies suggest that ATP is not directly involved in inflammatory cell chemotaxis, but rather is a prerequisite for the production of chemotactic factors at the site of inflammation [12]. It is proposed that ATP activates stromal or resident inflammatory cells (very likely tissue macrophages) to generate an inflammatory microenvironment that produces the chemotactic gradient for the neutrophils. Other investigators also see ATP not as true chemotactic signal but rather as a signal of cell distress that in turn a) induces the release of chemotactic factors from nearby cells and b) modulates activation of receptors for chemotactic factors [23]. However, evidence obtained by other investigators suggest that ATP (and ADP) might function as a true chemotactic signal [24,25] and that pannexins might be involved in the generation of the ATP gradient generated by apoptosing cells [26]. Among other nucleotides, UTP has been shown to have a strong chemotactic activity against neutrophils and mast cells [27]. More recently, hematopoietic stem cells were shown to chemotact in response to UTP [28,29]. Receptors involved are the G-protein coupled P2Yrs, mainly P2Y2, P2Y6, P2Y12 and P2Y13. Thus, inflammatory cells respond to a graded nucleotide concentration by moving towards the gradient core. Once inflammatory cells have reached the center of the inflammatory site, it is likely that the elevated ATP levels function as a “stop signal”, probably acting at the P2X7 receptor.

Dendritic cells (DCs) offer an interesting example of immune cell responses to a nucleotide gradient. In fact, only immature DCs chemotact to nucleotides, while mature DCs, despite they express P2 receptors to the same level as immature DCs, are fully non-responsive [30]. This suggests that, once DCs have localized and captured the Ag and migrated to the lymph nodes, there is no need to remain sensitive to attraction generated by cell-released danger signals, in fact mature DCs should lose nucleotide-dependent chemotactic activity in order to migrate from the ATP-rich inflammatory site to the lymph nodes. Exposure to a graded ATP concentration moulds DC responses in multiple fashions. It is well known that after Ag capture DC start a complex differentiation process that leads to maturation and therefore to full ability to present Ag to lymphocyte and start an immune response. Depending on the type and amount of cytokine released and the co-stimulatory molecules expressed, DC will drive naïve CD⁴⁺ T lymphocyte differentiation towards a Th1, Th2, Th17 or Treg phenotype. Such a differentiation process is strongly affected by the inflammatory microenvironment and by the ATP concentration. It has been shown by several groups that incubation of DCs in the presence of micromolar ATP concentrations drives DC maturation towards a Th2 phenotype [31,32], and more recently that autocrine ATP release modulates the differentiation of Treg cells [33,34]. The ATP-rich microenvironment affects other key immune cell functions such as phagocytosis, phagosome-lysosome fusion and release of cytotoxic mediators. Long ago, high extracellular ATP levels were shown to inhibit phagocytosis in mouse macrophages [35]. The inhibitory effect of ATP on particle ingestion was confirmed in human monocytes and shown to be P2X7-dependent [36]. On the other hand, UDP acting at the P2Y6 receptor is a potent trigger of phagocytosis in microglia [37]. Extracellular ATP has further distinct effects downhill to phagocytosis as ATP-mediated P2X7 stimulation is a strong inducer of phagosome-lysosome fusion and subsequent killing of ingested microorganisms [38,39]. The microbicidal activity due to P2X7 stimulation had been originally assigned to its pro-apoptotic effect, however subsequent experiments suggest that facilitation of phagosome-lysosome fusion has itself a potent microbicidal activity by exposing the ingested microrgansim to lysosomal content. This view is further supported by the ability of extracellular ATP to cause the release of reactive oxygen species (ROS) via the mitochondria or NADPH oxidase activation.

An essential aspect of the immunomodulatory activity of ATP and P2 receptors is the cytokine-releasing activity. This was first described for IL-1B [40], but has been later showed for several other crucial cytokines [41–44]. Secretion of other cytokines, e.g. IL-12, on the other hand is inhibited by extracellular ATP [31,42]. It must be stressed that ATP effects are very much dose- and receptor subtype- dependent, as while at low doses ATP has mainly an immunosuppressive, tolerogenic, activity (presumably acting at P2Y1 and/or P2Y11 receptors, at high doses the effect is mainly pro-inflammatory, very likely acting at P2X7 [45]. Therefore, low ATP doses preferentially activate immunosuppressive or tolerogenic pathways, while high doses trigger pro-inflammatory pathways [46]. ATP exerts a potent modulatory activity also on chemokine secretion, inducing release of CCl22 and decreasing LPS-induced secretion of CXCL10 and CCL5. The net effect of this modulation of cytokine and chemokine release in DCs is on one hand to reduce Th1 cell differentiation and recruitment and on the other to drive Th2 cell differentiation. Paradoxically, immunosuppressive pathways may also be activated via the classical pro-inflammatory P2X7 receptor, as reported by Robson and co-workers. These authors showed that mice deficient of CD39, the main ecto-ATP/ADPase, are protected against Con-A-induced hepatonecrosis because failure to degrade ATP causes an accumulation of this nucleotide to a level sufficient to trigger P2X7-mediated apoptosis of NKT cells, the main effector of Con-A induced liver injury [47]. ATP might also down-modulate the immune response by up-regulating thrombospondin and indoleamine 2,3-dioxigenase, factors known to inhibit T cell proliferation and stimulate TGF-β release. The shift in DC responses (from immunosuppresion to immunostimulation) in presence of increasing concentrations of extracellular ATP is consistent with the view that while graded exposure to DAMPs may preferentially cause adaptation to the new condition rather than an overt defensive response, a brisk and large increase in DAMPs is more likely to cause inflammation [45].

4. P2 receptors (P2X7) as drug targets

Identification of P2 receptors mediating the pro-inflammatory effects of extracellular nucleotides (mainly ATP) is a crucial issue in purinergic signalling as this might lead to the development of novel anti-inflammatory drugs. In vitro and in vivo data point to P2Y2 and P2X7 as the main P2 subtypes mediating the pro-inflammatory effects of ATP. P2Y2 is most probably responsible for recruitment of neutrophils, dendritic cells, eosinophils and macrophages at inflammatory sites, but it may also participate in release of pro-inflammatory factors such as for example elastase IL-33 or MCP-1/CCL2.

Among P2 receptors, P2X7 has received special attention since its participation to inflammation is more extensive and more thoroughly characterized. In the hope to develop novel drugs for the treatment of inflammation more information have been gathered on the involvement of this receptor in the immune response than for the other P2 receptors. P2X7 belongs to the P2X receptor family of ATP-gated cation channels, but differs from the other P2X receptors in its C-terminus which is about 200 amino acid longer. Brief activation of P2X7 with extracellular ATP in its tetra-anionic form, ATP⁴⁻, opens cation-specific ion channels. Prolonged ligation of P2X7 results in the formation of non-selective membrane pores, permeable to molecules of molecular mass up to 900 Da, as shown experimentally by the uptake of fluorescent dyes. Depending on the cell type, P2X7 stimulation triggers opening of non-selective pores which allows cationic and anionic dyes uptake. Formation of the non-selective pore is dependent on the cytoplasmic C-terminal domain of P2X7. Prolonged P2X7 activation can lead to membrane blebbing and cell death by lysis/necrosis or apoptosis, depending on the cell type. The cytotoxic effect dependent on P2X7 (originally named P2Z) activation was first described in mouse lymphocytes [9] and tentatively thought to be implicated in T-cell dependent cytotoxicity, but this hypothesis never received strong experimental support, thus it is fair to say that physiological significance of cell death mediated by the ligation of P2X7 by ATP remains unclear. However, in the light of the recent demonstration that ATP may reach several hundred micromolar levels in the interstitial fluid, the cytotoxic effect mediated by P2X7 should be re-considered as it is not unlikely that part of the cell injurious effects of inflammation are indeed explained by the cytotoxic activity of ATP via the P2X7 receptor. In this respect, it also needs to be stressed that albeit two other P2 receptors have been implicated in cytotoxicity, P2X2 and P2X4, P2X7 expression is both necessary and sufficient to support ATP-mediated cytotoxicity in all cell models so far investigated.

More recent evidence suggest that tonic, low level, activation of P2X7 may in some conditions trigger growth or promote survival [48] via a combined effect on endoplasmic reticulum (ER) and mitochondria Ca²⁺ content, mitochondrial potential and oxidative phosphorylation, and NFATc1 activation. Growth-promoting effects of P2X7 are relevant in T lymphocyte proliferation where a functional P2X7 receptor is needed to start mitogenic activation and to support growth [33]. It is not known if P2X7 is also implicated in proliferation in other immune cells. Further complexity in the participation of P2X7 in the regulation of immune cell functions is added by the recent discovery that a shorter P2X7 natural splice variant (named P2X7B) lacking the C-terminal region is highly expressed in lymphocytes [49]. This short isoform possesses most properties of the longer isoform except for the non-selective pore formation. In a cell model (HEK293 cells) transfected with both receptors, the longer P2X7 isoform (P2X7A) and P2X7 B can assemble on the cell plasma membrane forming a heterotrimeric receptor with distinct functional properties. Surprisingly, the shorter P2X7 isoform, rather than acting as a dominant negative, stabilizes plasma membrane expression of P2X7 and potentiates its responses. Thus, P2X7 may have different properties depending upon the ratio of shorter vs. longer isoforms in the heterotrimer. It is unknown whether native P2X7A/B heterotrimers exist and if so what their function might be, but it is intriguing that mitogenic stimulation of T lymphocytes causes an increase in P2X7A, but most notably in P2X7B transcription [49]. Stimulaton of T cell receptors on T lymphocytes triggers the release of intracellular ATP which stimulates cell surface P2X7 leading to Ca²⁺ influx, NFAT activation and IL-2 synthesis.

There is growing evidence that P2X7 participates in CD⁴⁺ T lymphocytes differentiation in multiple and as yet only partially understood ways. One one hand P2X7 appears to be constitutively expressed by Treg cells, but on the other P2X7 stimulation inhibits Treg responses and skews their differentiation towards a Th17 phenotype [34]. A role for P2X7 has also been shown in macrophage differentiation where this receptor modulates cell fusion in the typical process of multinucleated giant cell formation occurring in granuloma, or during osteoclast differentiation in the bone [50]. These studies highlighted an additional function of P2X7 in purinergic signalling: a pathway for ATP release. In fact, until recently search for the elusive pathways mediating non-lytic ATP release had mainly concentrated on vesicle-mediated release or plasma membrane channels belonging to the connexin/pannexin family [20,51]. Now, we know that P2X7 can be a pathway for ATP release thus generating its own ligand. Experiments in osteoclasts further point to the existence on the plasma membrane of a structured complex of molecules involved in ATP signalling (ATP signalosome?) comprising ATP-generating systems (P2X7 and other plasma membrane conduits), P2 receptors, ecto-ATPases and P1 receptors. It is increasingly clear that the multiple facets of purinergic signalling and its profound pathophysiological implications can only be understood in the context of the “ATP signalosome”.

5. P2X7-dependent activation of proteolytic pathways

It has been shown by several investigators that P2X7 stimulation leads to the activation of various caspases or metalloproteases. A paradigm of this activity is the proteolytic processing and release of IL-1β and IL-18 by microglia and macrophages [52]. Schematically, the release of these interleukins requires two signals. The first signal via Toll-like receptors drives pro-IL-1β and pro-IL-18 expression and accumulation in the cytosol. The second signal via P2X7 triggers the proteolytic cleavage of leaderless pro-IL-1β and pro-IL-18 and the release of the mature cytokines. In fact, P2X7 is a potent activator of the NLRP3 inflammasome, which is a large multimeric protein platform composed of NLRP3, the adaptator ASC and procaspase 1. Oligomerisation of procaspase 1 leads to its proteolytic activation and the production of active caspase 1 which is involved in the proteolytic processing of pro-interleukins into their active form. P2X7 activation also triggers the proteolytic cleavage of plasma membrane proteins such as l-Selectin, CD23, TNFα, CD27, matrix metalloproteinase-9 and interleukin-6 receptor [53–57]. Recently Delarasse et al. have shown that P2X7 activation induces the proteolytic cleavage and shedding of the soluble fragment of the amyloid precursor protein (sAPPα) from neuroblastoma cells, in the absence of ADAM9, 10 and 17 [58]. The P2X7-dependent protease involved in the generation of sAPPα fragment was not identified but several pharmacological inhibitors suggested that a metalloprotease(s) is involved in this processing pathway [58]. Thus, P2X7 stimulation can stimulate other ADAMs than ADAM10 and 17.

6. P2X7 and autoimmune diseases

Macrophages from P2X7 ko mice are unable to release mature IL-1β and IL-18 after LPS stimulation followed by P2X7 activation [59]. Therefore, the potential role of P2X7 in systemic or organ-specific auto-immune diseases has been tested in mice deficient for the P2X7 gene. In a monoclonal anti-collagen induced arthritis, Labasi et al. have clearly shown that mice lacking P2X7 developed less severe arthritis than wt animals [60]. Less severe cartilage destructions and synovial inflammation as well as decrease in collagen cleavage products were found in P2X7-deficient mice suggesting that the lack of P2X7 leads to a decrease in pro-inflammatory cytokines such as IL-1β known to be involved in arthritis severity. In humans, Portales-Cervantes et al. have analysed the role of P2X7 in systemic lupus erythematosus (SLE) and rheumatoid arthritis (RA) [61]. They studied P2X7 expression and function and two genetic polymorphisms (1513 A/C and 762 T/C) in 101 SLE and 122 RA patients compared to healthy controls. They did not find differences in the frequency of the genetic polymorphisms in patients and controls. In contrast, Al-Shukaili et al. found that the presence of rheumatoid factor and anti-MCV autoantibody was significantly associated with the 1513 A/C polymorphism, in RA patients [62]. In RA patients, Portales-Cervantes et al. found a significant increase in IL-1β released from ATP-stimulated monocytes as compared to those of healthy controls. In contrast, ATP-triggered monocytes from SLE patients showed a significant decrease in IL-1β production as compared to controls. Thus, the observations made in RA patients are compatible with those made in the mouse model of arthritis even though the defects found in humans are moderate. Overall, P2X7 seems to be involved in arthritis severity while its role in SLE remains elusive. The potential involvement of P2X7 in SLE relies on genomic studies which identified 14 lupus susceptibility loci among which the human SLE locus SLEB4 at 12q24 which includes the gene encoding P2X7 [63,64].

Since the incidence and gravity of arthritis induced by anti-collagen antibodies is reduced in P2X7 ko mice, two different groups used this mouse strain to study the role of P2X7 in experimental autoimmune encephalomyelitis (EAE), an inflammatory demyelinating disease of the CNS induced in susceptible species and strains of mice by immunisation with CNS myelin, or myelin proteins. EAE shares many similarities with the various Multiple Sclerosis subtypes. Interestingly, Chen and Brosnan found that P2X7 deficient animals develop more severe EAE than wt animals [65]. Indeed, a significant increase in the number of lesions in the CNS was found in P2X7 ko as compared to wt animals. Importantly, using bone marrow chimeras, Chen and Brosnan showed that bone marrow derived cells from P2X7 ko mice increased the severity of the disease in irradiated wt mice. Furthermore, the number of apoptotic cells in brain and spinal cord was higher in wt CNS than in P2X7 ko animals. Overall, these data suggest that the decrease in apoptosis of lymphocytes in the CNS plays a major role in the worsening of EAE in P2X7 ko. In contrast to this study, Sharp et al. found, in a different P2X7 ko mouse strain, that the incidence of MOG-induced EAE is reduced when compared to wt animals [66]. The discrepancy between these studies might due to the usage of two different P2X7 ko lines: Pfizer vs GlaxoSmithKline (GSK) P2X7 ko animals. Indeed, it was recently suggested that the GSK-P2X7 ko mice express a functional P2X7 in T lymphocytes but not in macrophages and dendritic cells [67]. In addition, Nicke et al. have identified a P2X7 (k) isoform which is expressed in the GSK but not in the Pfizer P2X7 ko [68]. This P2X7 (k) variant was shown to have an 8-fold higher ligand sensitivity and to transduce signals more efficiently than the commonly expressed P2X7 (a) isoform. Thus, these results showing that the GSK P2X7 ko mice might express a functional isoform of P2X7 in T lymphocytes suggest that in these ko animals the decrease in the EAE incidence might be due to the presence of P2X7 (k) on T lymphocytes, which down modulates autoimmune responses.

It is well established that cell populations different from lymphocytes express P2X7 in the CNS. In particular, microglial cells also express P2X7 which, after activation, triggers the production, among several other inflammatory mediators, of 2-arachidonoylglycerol (2-AG) an endocannabinoid that activates neuronal CB1 receptors [69]. Stimulation of these receptors reduces glutamate release and inhibits excitotoxicity decreasing tissue damages. In addition, activation of CB2 receptors expressed by microglial cells and lymphocytes decreases the production and release of pro-inflammatory cytokines and free radicals. Witting et al. have also determined that stimulation of P2X7 during EAE leads to increased endocannabinoid production and reduces cellular destructions. They found that the brain amounts of endocannabinoid were not augmented in wt mice suffering of EAE and showing axonal damages as compared to control animals [69]. These observations show that cellular destruction occurring during EAE does not lead to increased production of neuroprotective endocannabinoids as found in various neuropathologies. In addition, in animals undergoing EAE, the production of endocannabinoid (2-AG) was significantly decreased in the CNS of P2X7 ko animals as compared to wt mice witting [69]. This reduction in 2-AG levels found in P2X7 ko mice correlated well with increased tissue destructions found in these animals compared to wt mice. It is worth noticing that in addition to microglial cells, astrocytes express P2X7 and are able to produce 2-AG albeit in much lower amounts as compared to microglia. However, being the astrocyte population larger than microglia, one can hypothesize that astrocyte production of 2-AG significantly contributes to neuroprotection. Thus, expression of P2X7 by microglial cells and astrocytes should protect from EAE even though endocannabinoid production is partially inhibited by EAE indepently of P2X7.

Oligodendrocytes which are the target of autoimmune attacks in EAE were shown to express functional P2X7 and triggering of this receptor lead to oligodendrocyte death in vitro and in vivo [70]. Interestingly, treatment of mice with pharmacological inhibitors of P2X7 inhibited chronic EAE by decreasing demyelination. The analyses of optic nerves from MS patients and healthy controls revealed that P2X7 mRNA and protein were significantly increased in patients compared to controls. Thus, P2X7 stimulation may increase tissue damage in the CNS of MS patients as it does in mice suffering from EAE. Altogether, these results suggest that P2X7 on oligodendrocytes aggravate EAE while its presence on aggressive T lymphocytes, microglial cells and astrocytes protects from this organ-specific autoimmune disease.

In four mouse models of inflammatory bowel disease Gulbransen et al. have demonstrated that enteric neurons which express P2X7 undergo neuronal death following colitis inflammation [71]. Inhibition of P2X7 with its pharmacological inhibitor o-ATP prevented myenteric neuronal death but did not block the pathological signs associated with colitis such as macroscopic damages and weight losses. In addition, neuronal death was shown to be caspase-dependent and blocked by the pan-caspase inhibitor Z-VAD. The authors suggested that enteric neuronal death was dependent on pannexin-1 (Panx1) because it was inhibited by the selective peptide inhibitor ¹⁰Panx and the pharmacological inhibitor probenicid. The authors suggested that panx-1 activation might be needed to activate the inflammasome and determined which components of this complex were required for neuronal death. Using two ko mouse strains, they showed that enteric neural death involved the protein Asc but not NLRP3. Importantly, they found that blocking panx1, caspases or the Asc pathway inhibited neuronal death but did not improve weight loss and macroscopic damage. Overall, these results may be therapeutically relevant because panx1 inhibition could protect enteric neurons during colitis and secure innervation of colonic muscle. However, in this study the direct involvement of P2X7 and panx1 in enteric neuronal death was mainly shown with the use of pharmacological inhibitors, P2X7 and/or panx1 ko animals should have been used to strengthen their conclusions.

Kawamura et al. have studied the role of P2X7 in Concanavalin-A induced autoimmune hepatitis [72]. It is known that ConA induced hepatitis is mediated by NKT cells because their elimination is protective. It has previously been shown that cell surface proteins from mouse T lymphocytes treated with NAD are ADP-ribosylated by the cell surface enzyme ADP-ribosyl transferase 2 (ART2) [73]. ADP-ribosylation of P2X7 induces its activation leading to Ca²⁺ influx, non selective pore formation and cell death. Among liver mononuclear cells, NKT and T lymphocytes but not NK cells express ADP-ribosyltranferase activity and are ADP-ribosylated on P2X7 provided that NAD is added [72]. Mice treated by NAD 2 h before ConA injection are protected from hepatitis because liver NKT cells are functionally inactivated. Importantly, when NAD was injected into mice 3 h after ConA, hepatitis severity was increased and 40% of the treated mice died [72]. The involvement of P2X7 in this response is cleary established because administration of NAD into ConA-stimulated wt mice induces important liver destruction while in P2X7 ko mice liver damage was strongly reduced. These experiments clearly show that P2X7 mediates opposite effects in ConA-induced autoimmune hepatitis. Inhibitory signals are delivered to naive NKT cells protecting from liver injury while stimulatory ones are given to activated NKT lymphocytes exacerbating autoimmune hepatitis.

Recently, the role of P2X7 in the development of type 1 diabetes in Non obese diabetic (NOD) mice was evaluated by Yi-Guang Chen et al. [74]. They compared the incidence of type 1 diabetes in NOD mice of three genotypes : P2X7 wt, P2X7^+/−, P2X7 ko. They found no significant differences between the three genotypes in both sexes. However, the same authors previously established that type 1 diabetes is accelerated in NOD mice lacking CD38. CD38 is an ectoenzyme which hydrolyses NAD and thus down modulates ADP-ribosylation of P2X7 by ART2 on T lymphocytes. Hence, the lack of CD38 leads to a decrease in P2X7 dependent NAD-induced cell death. In NOD.CD38 ko mice, the acceleration of type I diabetes was attributed to NAD-induced cell death of CD⁴⁺ invariant NKT cells and Foxp3 regulatory T lymphocytes [74]. Importantly, when the P2X7 deficiency was introduced in NOD. CD38 ko mice, they found that the lack of P2X7 abolished accelerated development of diabetes. Interestingly, Yi-Guang Chen et al. demonstrated that the numbers of CD⁴⁺ invariant NKT cells and Foxp3 regulatory T lymphocytes were restored in NOD double ko mice to the levels found in NOD mice [74]. Altogether, these results demonstrate that ADP-ribosylation of P2X7 triggers NAD-induced cell death of subpopulations of lymphocytes involved in the down modulation of type 1 diabetes in NOD mice.

The evaluation of P2X7 role in autoimmune diseases is complicated by its presence on various sub-populations of lymphocytes, on antigen presenting cells and macrophages and on cells which are the target of the autoimmune attack. In addition, the biological function of P2X7 on these cells varies, its stimulation can lead to cell death, cell proliferation, secretion of pro-inflammatory cytokines, production and release of neuroprotective endocannabinoids. Thus, the production and use of P2X7 conditional ko animals is required to improve and clarify our knowledge on the role of P2X7 in autoimmune diseases.

7. Conclusion

Although major progress has been achieved over the past ten years in the field of purinergic signalling several exciting areas of research remain to be investigated. The biochemical characterization of P2X7 heterotrimers with distinct isoform composition should bring important information on the function of P2X7 on different subpopulations of T lymphocytes. This should stimulate proteomic analyses to identify the protein partners of these receptors and bolster efforts to characterize the biochemical pathways stimulated by the mitogenic P2X7 vs those triggering cell death. Up to now, the P2X7 ko mouse lines available have been produced by methods in which the neomycin cassette has not been removed from the inactivated gene. Thus, the impact of this cassette on the P2X4 gene, closely linked to P2X7, has not been evaluated. In addition, the removal of P2X7 from all cells in which it is naturally expressed prevents detailed analyses of its function on defined sub-populations of cells. The production of conditional ko mice in which one can control where and/or when P2X7 is expressed would be of major interest to delineate its role in physiological or pathological conditions. The development of more selective pharmacological agonists and antagonists of P2X receptors is required to analyse biochemical purinergic pathways more precisely. In addition, some of these drugs might be useful as therapeutic agents to block or stimulate P2X receptors in various pathophysiological conditions.