The purinoreceptor P2X7 is expressed on subsets of T cells and mediates responses of these cells to extracellular nucleotides such as ATP or NAD+. We identified P2X7 as a molecule highly up-regulated on conventional CD8αβ+ and unconventional CD8αα+ T cells of the intestinal epithelium of mice. In contrast, CD8+ T cells derived from spleen, mesenteric lymph nodes, and liver expressed only marginal levels of P2X7. However, P2X7 was highly up-regulated on CD8+ T cells from spleen and lymph nodes when T cells were activated in the presence of retinoic acid. High P2X7 expression on intestinal CD8+ T cells as well as on CD8+ T cells incubated with retinoic acid resulted in enhanced sensitivity of cells to extracellular nucleotides. Both cell populations showed a high level of apoptosis following incubation with NAD+ and the ATP derivative 2′,3′-O-(benzoyl-4-benzoyl)-ATP, and injection of NAD+ caused selective in vivo depletion of intestinal CD8+ T cells. Following oral infection with Listeria monocytogenes, P2X7-deficient mice showed similar CD8+ T cell responses in the spleen, but enhanced responses in the intestinal mucosa, when compared with similarly treated wild-type control mice. Overall, our observations define P2X7 as a new regulatory element in the control of CD8+ T cell responses in the intestinal mucosa.
The intestinal mucosa provides an essential barrier for microbes and prevents their spreading from the intestinal lumen into deeper tissue sites. The immune system of the intestine is crucial for the maintenance of this barrier. It has the challenging task of protecting against food-borne pathogens while at the same time tolerating Ags from nutrients and commensal microbes. The mechanisms that allow this discrimination are only ill-defined and a major focus of research. Immune cells of the intestinal mucosa reside in GALT but are also widely scattered within the epithelium and the underlying lamina propria (1). The epithelium harbors a large number of CD8+ T cells which, according to the composition of their TCR and CD8 molecules, segregate into two major subclasses. One group is characterized by the expression of the CD8αα homodimer and the αβTCR or the γδTCR. These CD8αα+ T cells are not restricted by conventional MHC class I molecules and their function is currently only poorly understood. Potential functions include the response to pathogens and stressed or transformed cells but also the protection and maintenance of the epithelial cell barrier (2). The second group of epithelial CD8+ T cells expresses the αβTCR and the CD8αβ heterodimer. These “conventional” CD8αβ+ T cells recognize Ag in the context of MHC class I and resemble regular CD8+ T cells in other lymphoid and nonlymphoid tissues (2, 3).
To respond adequately to the complex tasks within the intestinal mucosa, CD8+ T cells adapt with phenotypical and functional changes to this environment. These alterations are also apparent in an “intestine-specific” gene expression profile (4, 5, 6). The intestine-specific phenotype is induced and maintained by local signals, such as enhanced TGF-β concentrations in the mucosa, but is also a consequence of priming in GALT (2, 7). Priming in GALT leads to up-regulation of the chemokine receptor CCR9 and the α4β7 integrin, both important for migration of T cells to the intestinal mucosa (2, 8, 9). Recently, retinoic acid has been identified as a central inductor of CCR9 and α4β7 integrin in GALT (10). Dendritic cells patrolling the intestinal mucosa are able to oxidize inactive retinol to active retinoic acid. After migration to GALT, these dendritic cells provide the retinoic acid signal and thereby imprint the intestine phenotype during priming of T cells (10, 11). Beside CCR9 and α4β7 integrin, a growing number of genes with altered expression in mucosal T cells are currently being identified and most likely reflect part of the intestine-specific imprinting program (4, 5).
P2X7 is a purinergic receptor found on a large variety of cells, including cells of hematopoietic origin (12). In the mouse, P2X7 is triggered by extracellular ATP and NAD+. Whereas ATP directly binds and activates P2X7, stimulation by NAD+ requires the enzymatic activity of the ecto-ADP-ribosyltransferase 2.2 (ART2.2)3 (12, 13). ART2.2 transfers the ADP-ribose moiety from NAD+ to arginine residues on adjacent proteins. In the case of P2X7, ADP-ribosylation occurs at a site close to the ATP binding site and causes constitutive P2X7 activation (14). Depending on the cell type, stimulation of P2X7 can induce different molecular and cellular responses. In T cells, P2X7 stimulation initiates the opening of the P2X7-intrinsic nonselective ion channel which allows a rapid calcium influx (15). Subsequent responses are the surface exposure of phosphatidylserine (PS), proteolytic shedding of different surface proteins, and opening of a large membrane pore for hydrophilic substances up to 900 Da (15, 16). Prolonged stimulation eventually results in apoptosis and lysis of T cells, a process that has been termed nucleotide-induced cell death (15). In macrophages, triggering of P2X7 provides a cosignal for the activation of the IL-1 inflammasome and is an essential mediator of the synergism between ATP and TLR ligands in the maturation and release of bioactive IL-1β (12, 17).
As central elements of energy metabolism, ATP and NAD+ are present in millimolar concentrations in the cytoplasm of all cells. In contrast, only submicromolar concentrations are usually found in the extracellular milieu (18). Following inflammation, cell stress, or cell damage, both nucleotides are released and could then provide a danger signal for immune cells to sense inflammation and tissue destruction (18). In line with this assumption, a role for P2X7 in the induction of inflammatory responses has been proposed in different models for inflammatory and autoimmune diseases (12, 15). Induction of T cell apoptosis is also observed in mice in response to injection of NAD+ and less pronounced to injection of ATP. In BALB/c mice and particularly in CD38-deficient BALB/c mice, which are impaired in the degradation of extracellular NAD+, treatment with NAD+ causes massive depletion of all T cell subsets (18). C57BL/6 mice express a P2X7 allele with lower efficacy in the induction of T cell apoptosis (19). In these mice, depletion following NAD+ injection is restricted to T cell subsets with high P2X7 expression, such as regulatory T cells or NKT cells (20, 21).
In the current study, we identified P2X7 as a molecule highly up-regulated on conventional CD8αβ+ and unconventional CD8αα+ T cells of the intestinal epithelium. Peripheral CD8+ T cells showed only marginal P2X7 levels on their surface but expression was induced by retinoic acid. P2X7 was functional, since CD8+ T cells isolated from the intestinal epithelium as well as T cells activated in the presence of retinoic acid were highly sensitive to induction of apoptosis by NAD+ and the ATP derivative 2′,3′-O-(benzoyl-4-benzoyl)-ADP (BzATP) in vitro, and injection of NAD+ caused partial depletion of intestinal CD8+ T cells in vivo. A role for P2X7 in the regulation of intestinal T cell responses was also observed in P2X7-deficient mice. When compared with wild-type mice, infection of P2X7-deficient mice with Listeria monocytogenes showed similar CD8+ T cell responses in the spleen, but enhanced responses in the intestinal mucosa. In summary, our results define P2X7 as a new regulatory element for T cell responses in the intestinal mucosa.
Materials and Methods
Abs and reagents
Anti-CD16/CD32 mAb (clone: 2.4G2), anti-CD8α mAb (YTS169), anti-CD8β mAb (H35-17.2), anti-CD4 mAb (YTS191.1), anti-CD3 mAb (145-2C11), anti-CD28 mAb (37.51), anti-P2X7 mAb (Hano44) (22), anti-P2X7 pAb (K1G) (22), anti-ART2.2 mAb (Nika 102) (23), and anti-IFN-γ mAb (XMG1.2) were purified from serum (K1G) or hybridoma supernatants and Cy5-, Alexa Fluor 488-, or FITC-conjugated according to standard protocols. PE-conjugated anti-CD4 mAb (GK1.5), anti-CD8α mAb (53-6.7), anti-CD8β mAb (53-5.8) and anti-α4β7 integrin/LPAM-1 mAb (DATK32) were purchased from BD Pharmingen. H-2Kb/OVA257–264 tetramers were generated as previously described (24).
Mice and animal experiments
C57BL/6 mice, P2X7-deficient mice (25), and ART2.2-deficient mice (26) were bred at the central animal facility of the University Medical Center Hamburg-Eppendorf. All animal experiments were conducted according to the German animal protection law. Mice received 10 mg of NAD+ in 200 μl of PBS via injection into the lateral tail vein. All nucleotides were purchased from Sigma-Aldrich.
Mice were infected with L. monocytogenes strain EGD or with a L. monocytogenes strain recombinant for a secreted form of OVA (LmOVA) (27). For intragastric infection, L. monocytogenes was grown overnight in tryptic soy broth and washed twice in PBS. Bacterial density was determined by absorption at 600 nm, and bacteria were appropriately diluted in PBS (an OD600 value of 1 is equivalent to 109 bacteria/ml). Bacteria were administered in 200 μl of PBS by gastric intubation. The bacterial inoculum was always controlled by plating serial dilutions on tryptic soy broth agar plates. For determination of bacterial burdens in organs, mice were killed, organs were homogenized in PBS, serial dilutions of homogenates were plated on polymyxin-acriflavine-LiCl-ceftazidime-aesculin-mannitol-Listeria-selective agar supplemented with selective antibiotics (Merck), and colonies were counted after 48 h of incubation at 30°C.
Purification and staining of cells
Lymphocytes from spleen, mesenteric lymph nodes (MLN), liver, and the epithelium of small intestine, cecum, and colon were isolated as previously described (28). Cells were washed and incubated for 10 min with rat serum and anti-CD16/CD32 mAb to block unspecific Ab binding. Subsequently, cells were stained with Abs as indicated. After 30 min, cells were washed with PBS and fixed with PBS and 1% paraformaldehyde. For tetramer staining, 2 × 106 cells were incubated for 15 min at 4°C with rat serum, anti-CD16/CD32 mAb, and streptavidin (Molecular Probes) in PBS, 0.5% BSA, and 0.01% sodium azide. After incubation, cells were stained for 60 min at 4°C with the indicated Abs and streptavidin-PE-conjugated OVA257–264 tetramers. Subsequently, cells were washed with PBS, 0.5% BSA, and 0.01% sodium azide and diluted in PBS. Propidium iodide (PI) or 7-aminoactinomycin D was added before flow cytometry analysis. Cells were analyzed using a FACSCalibur and CellQuest 3.0 software or a FACSCanto and DIVA software (BD Biosciences).
Induction of cytokines and intracellular cytokine staining
Cells were cultured in a volume of 1 ml of RPMI 1640 medium supplemented with glutamine, 2-ME, gentamicin, and 10% heat-inactivated FCS (complete medium). Cells were stimulated for 4 h with 10−6 M of the peptide OVA aa 257–264 (OVA257–264, SIINFEKL). During the final 3.5 h of culture, 10 μg/ml brefeldin A (Sigma-Aldrich) was added. Cultured cells were washed and incubated for 5 min with rat serum and anti-CD16/CD32 mAb to block unspecific Ab binding. Subsequently, cells were stained with Abs against surface proteins and, after 30 min on ice, cells were washed with PBS and fixed for 20 min at room temperature with PBS and 4% paraformaldehyde. Cells were washed with PBS and 0.2% BSA, permeabilized with PBS, 0.1% BSA, and 0.3% saponin (Sigma-Aldrich), and incubated in this buffer with rat serum and anti-CD16/CD32 mAb. After 5 min, FITC- or Cy5-conjugated anti-IFN-γ mAb was added. After another 20 min on ice, cells were washed with PBS and fixed with PBS and 1% paraformaldehyde.
Annexin V staining
Lymphocytes from spleen and the epithelium of the small intestine were washed with NaCl buffer (0.14 M NaCl, 5 mM KCl, 0.01 M HEPES, and 0.01 M glucose) and incubated with the indicated concentrations of NAD+ or BzATP for 1 h at 37°C in NaCl buffer. Subsequently, cells were washed twice with Annexin V buffer (0.14 M NaCl, 2.5 mM CaCl2, and 0.01 M HEPES (pH 7.4)) and stained with Cy5-conjugated anti-CD8α mAb, FITC-conjugated annexin V (BD Pharmingen) and PI in Annexin V buffer for 20 min on ice. Immediately after the incubation, annexin V staining and PI uptake were analyzed by flow cytometry.
In vitro stimulation of T cells
CD8+ T cells from spleen and peripheral lymph nodes of naive mice were MACS purified by positive selection with anti-CD8 magnetic beads according to the manufacturer’s protocol (Miltenyi Biotec). Cells (2 × 106/ml) were incubated in complete medium in 24-well plates that had been coated overnight with anti-CD3 mAb (1 μg/ml PBS) and anti-CD28 mAb (1 μg/ml PBS). After 2 days, cells were washed and incubated in the 3-fold volume of complete medium containing 20 U/ml recombinant murine IL-2 (Roche). Where indicated, cultures contained in addition 10 nM all-trans retinoic acid (Sigma-Aldrich), 5 ng/ml TGF-β (R&D Systems), or a combination of both. After a further 3 days, cells were stained with Abs as indicated.
Unless otherwise stated, experimental groups consisted of at least three mice per group and mice were independently analyzed. For statistical analysis of cell frequencies and cell numbers, we applied the Student t test. Bacterial titers were compared using the Mann-Whitney U test. Differences were considered significant with p < 0.05.
Intraepithelial CD8+ lymphocytes express high levels of P2X7
In an approach to identify proteins involved in the regulation of intestinal CD8+ T cells, we could identify P2X7 as a molecule highly up-regulated on intestinal CD8+ T cells. In C57BL/6 mice, CD8+ T cells isolated from spleen, liver, and MLN showed only a low level of P2X7 surface expression (Fig. 1). In contrast, the majority of CD8+ T cells from the epithelium of small intestine, cecum, and colon displayed strong P2X7 expression. In these tissues, virtually all unconventional CD8αα+ T cells were highly positive for P2X7 as detected by mAb Hano44 (Fig. 1) and pAb K1G (data not shown). Conventional CD8+ T cells characterized by the expression of CD8β were divided into two distinct populations with high and low P2X7 expression, respectively. Whereas in the small intestine epithelium, CD8αβ+ T cells were mainly P2X7high, the epithelium of cecum and colon contained also major populations of CD8αβ+ T cells with low P2X7 expression.
ART2.2 is the major ADP-ribosyltransferase on T cells from C57BL/6 mice and is responsible for the activation of P2X7 by NAD+-dependent ADP-ribosylation (13, 14). Therefore, ART2.2 expression on CD8+ T cells was determined (Fig. 1). In contrast to P2X7, ART2.2 was detected on the majority of conventional CD8αβ+ T cells independent of the tissue of isolation as well as on most unconventional CD8αα+ T cells from the intestinal epithelium.
Intraepithelial CD8+ lymphocytes are sensitive to induction of apoptosis by nucleotides
Stimulation of P2X7 with ATP or ADP-ribosylation of P2X7 mediated by ART2.2 causes rapid PS exposure on the T cell surface and subsequently membrane disintegration and apoptosis of T cells (13). To determine whether intraepithelial CD8+ T cells were sensitive to nucleotides, lymphocytes were isolated from spleen and small intestinal epithelium and incubated for 60 min with different concentrations of NAD+ or BzATP. The ATP analog BzATP is a more potent agonist for P2X7 than ATP (29). PS exposure and loss of membrane integrity were determined by annexin V staining and PI uptake, respectively (Fig. 2). As has been demonstrated before (18, 20, 30), CD8+ T cells derived from the spleen of C57BL/6 mice were relatively resistant to BzATP or NAD+ incubation. In contrast, CD8+ T cells derived from the small intestinal epithelium showed PS exposure and PI uptake following culture with BzATP and NAD+. For both nucleotides, the response was mediated by P2X7 since T cells isolated from spleen and intestinal epithelium of P2X7-deficient C57BL/6 mice proved to be resistant to NAD+ and BzATP (Fig. 2).
Retinoic acid causes up-regulation of P2X7 on purified CD8+ T cells
Retinoic acid as well as TGF-β are considered as central factors involved in the imprinting of an intestine-specific T cell phenotype (4). To test whether retinoic acid or TGF-β influences the expression levels of P2X7, CD8+ T cells purified from spleen and peripheral lymph nodes (excluding MLN) were stimulated with anti-CD3 mAb and anti-CD28 mAb and, after 2 days, were further incubated with all-trans retinoic acid, TGF-β, or a combination of both. After another 3 days, the expression levels of the α4β7 integrin, ART2.2, and P2X7 were analyzed (Fig. 3). α4β7 integrin showed a modest expression level on activated CD8+ T cells. The addition of retinoic acid significantly augmented the surface expression of α4β7 integrin. TGF-β alone caused no change in α4β7 integrin expression and, when combined with retinoic acid, TGF-β reduced induction of α4β7 integrin expression. There was only marginal expression of P2X7 on activated CD8+ T cells. Similar to the α4β7 integrin, the addition of retinoic acid to cultures induced strong up-regulation of P2X7 on CD8+ T cells. TGF-β neither changed the marginal P2X7 expression on cells cultured without retinoic acid nor the high expression induced by retinoic acid. The addition of retinoic acid enhanced the moderate expression of ART2.2 on activated CD8+ T cells while TGF-β showed little, if any, effect.
Retinoic acid induces sensitivity against extracellular nucleotides in CD8+ T cells
To test whether the up-regulation of P2X7 following culture with retinoic acid was accompanied by increased sensitivity to extracellular nucleotides, activated CD8+ T cells were incubated for 60 min with BzATP or NAD+ and PS exposure and PI uptake were determined (Fig. 4). Consistent with the low P2X7 expression, CD8+ T cells activated for 5 days were relatively resistant against induction of apoptosis by both BzATP and NAD+. Upon incubation with retinoic acid, activated CD8+ T cells acquired sensitivity to BzATP as well as to NAD+. Sensitivity to nucleotides was mediated by ART2.2 and P2X7 since no response to NAD+ or BzATP and NAD+ could be detected when CD8+ T cells were derived from mice deficient in ART2.2 or P2X7, respectively (data not shown).
Intraepithelial CD8+ lymphocytes are sensitive to NAD+ in vivo
In C57BL/6 mice, injection of NAD+ causes only depletion of T cell subsets with high P2X7 expression (20, 21). To test whether the enhanced expression of P2X7 on intraepithelial CD8+ T cells was accompanied by enhanced sensitivity to nucleotides in vivo, C57BL/6 mice received 10 mg of NAD+ i.v (18). Twenty hours later, lymphocytes from spleen, MLN, and small intestinal epithelium were purified and analyzed by flow cytometry (Fig. 5). As expected for C57BL/6 mice, injection of NAD+ caused no or only minor changes in the frequencies and total number of CD4+ and CD8+ T cells in spleen and MLN. In contrast, the frequencies and numbers of CD8+ T cells in the small intestinal epithelium were significantly reduced. Among CD8+ T cells, depletion affected mainly CD8αα+ T cells. In P2X7-deficient mice, depletion of intraepithelial CD8αα+ T cells following NAD+ injection was not observed, confirming the central role of P2X7 in NAD-induced T cell depletion (data not shown).
Intestinal CD8+ T cells generated in response to oral L. monocytogenes infection express high P2X7 levels
Oral L. monocytogenes infection causes priming of CD8+ T cells in GALT. To analyze the expression of P2X7 on CD8+ T cells generated during oral infection, mice orally received bacteria of a L. monocytogenes strain recombinant for OVA (LmOVA). Infection with LmOVA induces a strong response of conventional CD8αβ+ T cells against the immunodominant OVA peptide OVA257–264 (27). Nine days after oral infection, lymphocytes were isolated from spleen, liver, MLN, and the small intestinal epithelium, and OVA257–264-restricted CD8+ T cells were visualized with OVA257–264 tetramers (Fig. 6). OVA257–264 tetramer-positive CD8+ T cells derived from spleen, MLN, or liver displayed only moderate levels of P2X7 expression. In contrast, OVA257–264 tetramer-positive CD8+ T cells isolated from the epithelium of the small intestine showed highly elevated surface expression of P2X7. This observation is consistent with results from expression-profiling experiments with two different sets of microarrays (4), showing a 4.5-fold or 13.4-fold increase in the expression of P2X7 mRNA in OVA257–264 tetramer+CD8+ T cells isolated from small intestinal epithelium vs spleen of mice orally infected with LmOVA. The high P2X7 expression on specific OVA257–264-CD8+ T cells is also consistent with the high expression level of P2X7 on conventional CD8αβ+ T cells from the epithelium of naive mice. Furthermore, the intestine-restricted P2X7 expression indicates that up-regulation of P2X7 is not a general process following activation and differentiation of CD8+ T cells in response to infection but rather a consequence of T cell priming or T cell localization in the intestinal mucosa.
P2X7 expression on intestinal OVA257–264-specific CD8+ T cells also confers sensitivity to extracellular NAD+ to these cells (Fig. 7). C57BL/6 mice were orally infected with LmOVA. After 8 days, one group of infected mice was treated with 10 mg of NAD+ and 1 day later the response to NAD+ was analyzed. In this set of experiments, we used production of IFN-γ in response to peptide stimulation as means to visualize OVA257–254-specific CD8αβ+ T cells. Consistent with results from the tetramer analyses (Fig. 6), spleen-derived OVA257–264-specific CD8+ T cells, as determined by IFN-γ production, expressed little, if any, P2X7. In contrast, the majority of OVA257–264-specific CD8+ T cells isolated from the intestinal epithelium displayed enhanced surface levels of P2X7. Injection of NAD+ did not affect frequencies and numbers of OVA257–264-specific CD8+ T cells in spleens of LmOVA-infected mice. In the intestinal epithelium, the frequencies and numbers of OVA257–264-specific CD8αβ+ T cells were reduced after NAD+ administration, and the remaining OVA257–264-specific CD8αβ+ T cells expressed only low surface levels of P2X7. Thus, P2X7 expression acquired during priming or localization in the intestinal mucosa is connected to enhanced sensitivity of cells to extracellular nucleotides in vivo.
P2X7-deficient mice demonstrate enhanced intestinal CD8+ T cell responses to infection with L. monocytogenes
Because CD8+ T cells of the intestinal epithelium generated during L. monocytogenes infection showed enhanced P2X7 expression, we tested whether the absence of P2X7 alters the intestinal T cell response to this pathogen. In a first set of experiments, P2X7-deficient and control mice were orally infected with L. monocytogenes and bacterial titers in spleen, MLN, liver, and intestine, including tissue and luminal content of the small intestine, cecum, and colon, were determined on different days postinfection. For these experiments, we applied the wild-type L. monocytogenes strain EGD which is ∼10-fold more virulent than the recombinant LmOVA strain (our unpublished observation). Fig. 8 shows results for bacterial titers for days 2, 5, 7, and 9 after intragastric infection with 109 bacteria. Most likely as a consequence of the intragastric infection route, we observed a relatively high variation of bacterial titers between individual mice of both experimental groups. P2X7-deficient mice showed slightly higher titers in spleen and liver on the early days of infection; however, differences never reached a statistically significant level. Overall, deficiency in P2X7 did not cause gross alterations in susceptibility to L. monocytogenes infection. Mutant and wild-type mice controlled infection and most mice of both groups had eradicated bacteria by day 9 after infection.
In a parallel set of experiments, P2X7-deficient and control mice were orally infected with LmOVA. Cells isolated from spleen and small intestinal mucosa were incubated with OVA257–264 and the frequencies and numbers of OVA257–264-specific CD8+ T cells were determined by intracellular IFN-γ staining (Fig. 9). Both mouse strains were able to mount CD8+ T cell responses against the immunodominant epitope OVA257–264, and spleens of both mouse strains displayed similar numbers of specific CD8+ T cells. In contrast, the small intestinal epithelium of P2X7-deficient mice contained significantly elevated numbers of OVA257–264-specific CD8+ T cells. Thus, the absence of P2X7 led to enhanced CD8+ T cell responses in the intestinal mucosa.
P2X7 was highly expressed on the majority of CD8+ T cells isolated from the epithelium of the small intestine, cecum, and colon of naive mice. Basically, all unconventional CD8αα+ T cells and large subsets of conventional CD8αβ+ T cells showed high levels of P2X7 expression in these tissues but not in any peripheral tissue analyzed. Results from L. monocytogenes infection studies further suggest that for conventional CD8αβ+ T cells, this expression profile is not a consequence of the activation status but rather reflects the location in the intestinal epithelium and/or priming in GALT. T cells with the same specificity and from the same stage of infection and immune response were P2X7low in nonintestinal tissues and largely P2X7high in the small intestinal epithelium.
The origin of the P2X7lowCD8αβ+ T cell subset in the intestinal epithelium is currently unclear. We and others could demonstrate that conventional CD8αβ+ effector and memory T cells primed in central secondary lymphatic tissues can enter the intestinal mucosa (31, 32, 33). Thus, P2X7lowCD8αβ+ T cells could represent recent immigrants derived from nonintestinal sites. Initial analyses of i.v. infected mice indeed showed accumulation of P2X7low L. monocytogenes-specific T cells in the intestinal epithelium (our unpublished observation). This result also points to the site of priming as the main determinant of P2X7 expression on CD8αβ+ T cells and supports the concept of P2X7 up-regulation as a consequence of priming in GALT.
Incubation of activated conventional CD8+ T cells with all-trans retinoic acid induced up-regulation of P2X7 and ART2.2 expression. In contrast, TGF-β alone or in combination with retinoic acid had only marginal effects on the expression of these molecules. Together with the intestine-restricted expression of P2X7 on CD8+ T cells, this result suggests that up-regulation of P2X7 represents part of the intestine-specific imprinting program during priming by mucosa-derived dendritic cells, similar to the induction of CCR9 or α4β7 integrins and the down-modulation of NK receptors (4, 5, 10). Interestingly, our results resemble observations for regulatory T (Treg) cells. Treg cells express high surface levels of P2X7 and are particularly sensitive to extracellular nucleotides (20). Furthermore, inducible Treg cells are generated following activation and culture with retinoic acid and TGF-β (34, 35, 36). Whether P2X7high CD8 T cells possess regulatory functions or whether P2X7 expression is in any way linked to a suppressive function of T cells is unclear and currently under investigation.
CD8+ T cells isolated from spleens or lymph nodes of C57BL/6 mice have been shown to be relatively resistant to extracellular nucleotides (13, 20, 30). Furthermore, in vitro activation of T cells causes down-modulation of P2X7 and leads to loss of sensitivity to nucleotides (13). CD8+ T cells isolated from the intestinal epithelium or activated splenic CD8+ T cells cultured in the presence of retinoic acid showed a different behavior. These cells responded to both NAD+ and BzATP with rapid PS exposure and PI uptake, despite expression of the relatively resistant P2X7 allele of C57BL/6 mice (19, 37). Thus, a high expression level of P2X7 appears to compensate for the low sensitivity of the P2X7 variant of C57BL/6 mice. In this respect, P2X7highCD8+ T cells resemble Treg cells of C57BL6 mice, which express high levels of P2X7 and display high sensitivity to extracellular nucleotides as well (20). P2X7highCD8+ T cells were also sensitive to extracellular nucleotides in vivo. Injection of NAD+ caused massive depletion of CD8αα+ T cells in the intestinal epithelium of naive mice and, in mice orally infected with LmOVA, NAD+ treatment resulted in a loss of P2X7high OVA257–264-specific conventional CD8+ T cells in this tissue site. Remaining cells displayed only low P2X7 expression and could reflect down-modulation of P2X7 on mucosa-resident conventional CD8αβ+ T cells or recruitment of P2X7low cells from peripheral tissue sites into the mucosa. These results not only demonstrate that P2X7 is functional on intestinal CD8+ T cells, they also indicate that the machinery translating the P2X7 signal into an apoptotic outcome is present and active in these cells. Thus, the up-regulation of P2X7 on intestinal CD8+ T cells appears to be part of a differentiation program allowing the effective response of these cells to extracellular nucleotides. This notion is further supported by the coexpression of high ART2.2 surface levels on intestine-derived CD8+ T cells or on CD8+ T cells cultured with retinoic acid. Coexpression of ART2.2 with P2X7 renders CD8+ T cells sensitive to extracellular NAD+, which due to the lower concentrations of NAD+ required for P2X7 stimulation might be the more relevant P2X7 trigger for these cells when compared with extracellular ATP (13).
The restricted P2X7 expression and sensitivity to extracellular nucleotides imply a particular function for P2X7 in the intestinal epithelium. In an attempt to characterize this function, the immune response of P2X7-deficient mice orally infected with L. monocytogenes was analyzed. Due to the intragastric infection route, we observed a high degree of variation in bacterial titers between individual mice. However, all P2X7-deficient mice could restrict L. monocytogenes replication and at day 9 after infection, most P2X7-deficient mice had cleared the infection. Thus, P2X7-deficient mice were not severely impaired in their innate and acquired response to L. monocytogenes when compared with wild-type control mice. P2X7 has been postulated as an important component in the induction and maturation of IL-1β and IL-18 during inflammation (38). Furthermore, P2X7 deficiency caused reduced production of IL-6, TNF-α, and IFN-γ in different inflammation models, most likely as a consequence of impaired release of mature IL-1β and IL-18 (25, 38, 39, 40). Proinflammatory cytokines, particularly IL-6, TNF-α, and IFN-γ, are crucial elements of the immune response against the intracellular pathogen L. monocytogenes (41). However, our infection studies indicate that the immune response against L. monocytogenes is largely independent of P2X7. Given the crucial role of the inflammatory cytokines, the result also implies that these cytokines are produced in sufficient amounts to allow the control of infection. Control of L. monocytogenes in P2X7-deficient mice is not completely unexpected. As a main immune escape strategy, L. monocytogenes damages the phagosome membrane and migrates into the cytoplasm (42). In contrast to extracellular bacteria or to intracellular bacteria remaining in phagolysosomal compartments, the cytosolic location of L. monocytogenes allows effective activation of the IL-1β inflammasome via an alternative pathway, which is largely independent of extracellular ATP and P2X7 (43). Furthermore, infection studies with other intracellular bacteria revealed efficient bacterial clearance in P2X7-deficient mice (44, 45). P2X7-deficient mice showed slightly enhanced bacteria titers in the lower genital tract upon vaginal infection with Chlamydia muridarum; however, clearance of bacteria was similar to that of control mice (44). In vitro infection studies proposed a significant role of P2X7 in the control of mycobacteria and genetic linkage studies connect certain human P2X7 alleles with enhanced susceptibility to Mycobacterium tuberculosis (46). However, upon aerosol infection, P2X7-deficient mice demonstrated normal control of M. tuberculosis when compared with control mice (45). In summary, the role of P2X7 and, more generally, the role of extracellular nucleotides for the induction and progress of immune responses against infection are still unclear and afford further detailed analysis.
When OVA257–267-specific CD8+ T cell responses induced during infection with recombinant L. monocytogenes were analyzed, P2X7-deficient mice showed normal responses in the spleen, but enhanced responses in the intestinal epithelium. This enhanced response was not due to general alterations in the cellularity of the intestinal mucosa because naive as well as L. monocytogenes-infected P2X7-deficient mice displayed similar frequencies and total numbers of major T cell populations (CD4+, CD8αα+, and CD8αβ+ T cells) in spleen and intestinal epithelium when compared with respective control mice (our unpublished observation). Enhanced T cell responses in P2X7-deficient mice have been described before in infection and inflammation models (40, 45). P2X7-deficient mice infected with M. tuberculosis showed enhanced CD4+ T cell responses in infected lungs, and following induction of experimental autoimmune encephalomyelitis, T cells from P2X7-deficient mice proliferated more vigorously in response to stimulation with the autoantigen myelin oligodendrocyte glycoprotein (40, 45). In the latter case, unresponsiveness of T cells to extracellular nucleotides generated during infection or inflammation was suggested as possible reason for enhanced T cell responses. Likewise, nucleotides released at the site of L. monocytogenes infection, either by direct tissue damage or following inflammation could restrict the response of P2X7highCD8+ T cells in the intestinal mucosa. Deficiency of P2X7 should consequently result in enhanced intestinal CD8+ T responses. However, there are alternative explanations. P2X7 is expressed on other hematopoietic and nonhematopoietic cells and absence of P2X7 from these cells could be responsible for alterations of the immune response to L. monocytogenes. The absence of P2X7 could alter the expression of stimulatory or inhibitory cytokines, or the expression of surface ligands involved in the control of CD8+ T cell priming and proliferation. Finally, lack of P2X7 expression could cause changes in T cell migration and enhance accumulation of specific CD8+ T cells in the intestinal epithelium. Currently, our results do not allow us to prefer or exclude any of these mechanisms.
In summary, our results demonstrate that intestinal CD8+ T cells express high levels of P2X7 and are particularly sensitive to extracellular nucleotides. High P2X7 expression is largely independent of the activation status of cells and most likely reflects part of the gene expression profile imprinted on CD8+ T cells during priming in GALT. Finally, our results suggest that P2X7 is part of a so far unrecognized regulatory mechanism for intestinal CD8+ T cells.
We thank Fabienne Seyfried and Marion Nissen for technical assistance and Dr. Chris Gabel (Pfizer Inc.) for providing P2X7-deficient mice.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
F.K.-N. and F.H. were supported by the Deutsche Forschungsgemeinschaft (DFG NO310/6-2).
Abbreviations used in this paper: ART, ADP-ribosyltransferase; BzATP, 2′,3′-O-(benzoyl-4-benzoyl)-ATP; LmOVA, Listeria monocytogenes recombinant for ovalbumin; MLN, mesenteric lymph node; PI, propidium iodide; PS, phosphatidylserine; Treg, regulatory T.