P2X7 receptor-dependent tuning of gut epithelial responses to infection

Infection and injury of the gut are associated with cell damage and release of molecules such as extracellular adenosine 5′-triphosphate (ATP), which is recognised by the purinergic P2X7 receptor (P2X7R). P2X7R is widely expressed in the gut by antigen-presenting cells (APCs) and epithelial cells, but the role of the P2X7R on epithelial cells is poorly understood. We investigated P2X7R in intestinal epithelium in vitro and in vivo using two model infections, Toxoplasma gondii and Trichinella spiralis. Lipopolysaccharide and ATP treatment of intestinal epithelial cells and infection with T. gondii in vitro did not promote inflammasome-associated interleukin-1β (IL-1β) or IL-18 secretion, but promoted C–C motif chemokine ligand 5 (CCL5), tumour necrosis factor-α and IL-6 production that were significantly reduced when the P2X7R was blocked. Similarly, in vivo, infection with either T. spiralis or T. gondii induced rapid upregulation of epithelial CCL5 in wild-type (wild-type (WT)) mice that was significantly reduced in P2X7R−/− littermate controls. The effects of reduced epithelial CCL5 were assayed by investigating recruitment of dendritic cells (DCs) to the epithelium. Infection induced a rapid recruitment of CD11c+CD103+ DC subsets into the epithelial layer of WT mice but not P2X7R−/− mice. In vitro chemotaxis assays and bone marrow chimeras demonstrated the importance of epithelial P2X7R in DC recruitment. P2X7R signalling in epithelial cells mediates chemokine responses to promote initiation of host immunity to infection.

Infection and injury of the gut are associated with cell damage and release of molecules such as extracellular adenosine 5′-triphosphate (ATP), which is recognised by the purinergic P2X7 receptor (P2X7R). P2X7R is widely expressed in the gut by antigen-presenting cells (APCs) and epithelial cells, but the role of the P2X7R on epithelial cells is poorly understood. We investigated P2X7R in intestinal epithelium in vitro and in vivo using two model infections, Toxoplasma gondii and Trichinella spiralis. Lipopolysaccharide and ATP treatment of intestinal epithelial cells and infection with T. gondii in vitro did not promote inflammasome-associated interleukin-1β (IL-1β) or IL-18 secretion, but promoted C-C motif chemokine ligand 5 (CCL5), tumour necrosis factor-α and IL-6 production that were significantly reduced when the P2X7R was blocked. Similarly, in vivo, infection with either T. spiralis or T. gondii induced rapid upregulation of epithelial CCL5 in wild-type (wild-type (WT)) mice that was significantly reduced in P2X7R − / − littermate controls. The effects of reduced epithelial CCL5 were assayed by investigating recruitment of dendritic cells (DCs) to the epithelium. Infection induced a rapid recruitment of CD11c + CD103 + DC subsets into the epithelial layer of WT mice but not P2X7R − / − mice. In vitro chemotaxis assays and bone marrow chimeras demonstrated the importance of epithelial P2X7R in DC recruitment. P2X7R signalling in epithelial cells mediates chemokine responses to promote initiation of host immunity to infection. The gastrointestinal tract is a major route of entry for pathogens and is protected by a variety of defenses, the most important of which is the epithelial lining of the gastrointestinal tract. Luminal antigen recognition and processing begin at the epithelial level leading to orchestration of the coordinated activity of the underlying innate immune cells of the lamina propria including dendritic cells (DCs) and macrophages. As well as responding to the threat of invading pathogens, the gut must also be tolerant to host microbiota. Both host microbiota and pathogens express highly conserved molecular patterns that could potentially trigger host immunity, although most commensal microbiota are noninvasive. 1 Further, there is also physical separation of host microbiota and the epithelium in the gut because of the presence of a dense mucus layer. 1,2 Therefore, a key discriminating feature of a pathogen vs a host microbiota species is that pathogens actively invade and breach the epithelial layer causing cell damage and death.
Following tissue damage and necrotic cell death, irrespective of the initiating stimuli, a number of damage-associated molecular pattern molecules are released. Many damage-associated molecular pattern molecules are cytosolic or nuclear components including adenosine 5′-triphosphate (ATP). 3 Extracellular ATP is sensed by the ATP-gated purinergic P2X receptors (P2XRs). There are seven different mammalian genes encoding for P2X receptor subunits, P2X1-P2X7, and the P2X7R has a longer C-terminal sequence that is unique among the P2X family. 4,5 Activation of P2X7R by ATP induces Ca 2+ and Na + influx, K + efflux and causes an alteration of cell permeability by causing the formation of a large membrane pore, which eventually leads to cell death. 6 Under pathophysiological conditions, ATP released from dying cells can enhance P2X7R activation 7 and trigger activation of the NLRP3 ((NLR family, pyrin domain containing 3)) inflammasome-promoting secretion of the proinflammatory cytokines interleukin-1β (IL-1β) and IL-18. 4 Although the P2X7R is expressed by a variety of cell types including immune 4 and epithelial cells, 8 the activity of P2X7R is best described in immune cells.
P2X7R expression is linked with the development of visceral hypersensitivity in Trichinella spiralis (T. spiralis) infection. 9 P2X7R expression in immune cells is also implicated in the control of intracellular infections such as Toxoplasma gondii (T. gondii). Mutations in P2X7R are associated with enhanced susceptibility to T. gondii in humans. 10 Furthermore, P2X7R in macrophages and DCs promote phagolysosomal fusion and parasite killing in vitro. 10,11 However, little is known about the role of P2X7R in epithelial cells and the potential role of P2X7R as an epithelial sensor for infection and damage. Previously, we demonstrated that intestinal epithelial cells sense pathogens and initiate host immunity by secreting chemokines that promote rapid recruitment of DCs to the site of infection. 12 DCs reside in the gut as immature cells and in response to infection or injury are recruited from the gut and also from bone marrow-derived progenitors to the site of infection or injury. After priming they mature and express CCR7, which facilitates migration to secondary lymph nodes for the process of T-cell priming. 13 Thus, DCs are critical cells acting as the bridge between adaptive and innate immunity. Epithelial cell-derived inflammatory cytokines and chemokines are important for the initiation of gut inflammation and recruitment of inflammatory immune cells such DCs and other cells such as neutrophils. 14 Our data indicated that a deficient epithelial response is associated with enhanced susceptibility to infection. 12,15 Given that the intestinal epithelium is the major entry site of pathogens, and produces infection-induced chemokines and cytokines, 14 it seems probable that epithelial cells act as a sensor of infection.
In this study, we investigated the role of P2X7R in intestinal epithelial cells in response to two model infections, T. gondii and T. spiralis. In vitro we found that P2X7R in epithelia promoted chemokine production independently of inflammasome-associated cytokines IL-1β and IL-18. Investigations in vivo showed that P2X7R also promoted epithelial chemokine production in response to T. gondii and T. spiralis infection. Furthermore, reduced epithelial chemokine responses in P2X7R − / − mice were associated with a significant reduction in early infiltration of CD103 + DCs to the small intestinal epithelium. Our data indicate a novel role for P2X7R in epithelial cells in the initiation of small intestinal inflammation through chemokine production and recruitment of DCs to the site of infection.

P2X7R in epithelial cells promotes CCL5 production
Mouse colonic epithelial CMT-93 cells were infected with T. gondii plus or minus a selective P2X7R inhibitor, A-740003, for 24 h. Expression of P2X7R by CMT-93 cells was confirmed by flow cytometry (Figure 1a) and infection of epithelial cells was confirmed by immunohistochemistry and flow cytometry (Supplementary Figure 1). Secretion of ATP by intestinal epithelial cells in response to infection with T. gondii was confirmed using an ATP luciferinluciferase bioluminescence assay (Supplementary Figure 1). Infected epithelial cells produced tumour necrosis factor-α (TNF-α) and IL-6 and the levels of these cytokines were significantly decreased in the presence of the P2X7R inhibitor implicating P2X7R in the production of infection-induced cytokines (Figures 1b and c). As we have previously established that epithelial cells produce C-C motif chemokine ligand 5 (CCL5) in response to infection that drives DC recruitment, 12,15 we investigated production of CCL5. Infection with T. gondii induced a robust secretion of CCL5, which was significantly reduced by A-740003 ( Figure 1d, P = 0.0058). A-740003 treatment did not alter basal cytokine or chemokine levels ( Figure 1). These data suggest that P2X7R was specifically involved in the response to infection. P2X7R stimulation is known to be important for capase-1 activation and the assembly of the inflammasome in antigenpresenting cells. 16 Therefore, we tested whether P2X7R in epithelial cells also promoted production of the inflammasome-associated cytokines IL-1β and IL-18 in response to T. gondii infection. However, we found no detectable IL-1β production by CMT-93 cells (data not shown). IL-18 was secreted at low levels by CMT-93 cells, but this response was not affected by infection or P2X7R inhibition (Figure 1e). In contrast, we assayed the macrophage line THP-1 infected with T. gondii and showed they made a robust IL-1β response to infection (Supplementary Figure 2). Additional experiments were performed to contrast the role of P2X7R in epithelial vs macrophage lines. Immortalised bone marrow-derived mouse macrophages and macrophage vs epithelial cells were left unstimulated or stimulated with lipopolysaccharide (LPS) or LPS with ATP and IL-1β secretion was measured. The data showed that there was a significant induction of IL-1β in LPS-and ATP-treated macrophages but not IL-1β secretion by CMT-93 cells (Supplementary Figure 2). The possibility that the inhibitor was having a toxic effect on T. gondii was also investigated to ensure the reduction in chemokine production was not due to reduced parasite viability. T. gondii were pre-treated with A-740003 for 4 h, washed to remove all the inhibitor and then used to infect CMT-93 cells for 24 h. We found no difference in the percentage of parasite-infected epithelial cells if we used T. gondii that had been pre-treated with P2X7R inhibitor (data not shown). These data show that the P2X7R inhibitor is not toxic to the parasite. Collectively, our findings suggest that P2X7R-dependent regulation of chemokines and proinflammatory cytokine production in response to T. gondii infection is independent of inflammasome-regulated cytokines.
To investigate the role of the P2X7R further, we used a primary crypt organoid culture model. Small intestinal organoid cells from C57BL/6 (wild-type (WT)) or P2X7R − / − mice were treated with LPS. IL-1β and IL-18 were secreted at very low levels by primary small intestinal epithelial cells and this was not affected by LPS treatment (Figures 1f-h). A lack of P2X7R had no significant effect on IL-1β or IL-18 production in response to LPS, although there was a slight decrease in IL-18 production (Figures 1f and g). P2X7R deficiency did not influence basal levels of CCL5 secretion from crypt organoids (30 ± 6.8 pg ml − 1 (WT) vs 28 ± 6.0 pg ml − 1 (P2X7R − / − )). However, in agreement with our data using the epithelial cell line, WT organoids secreted significantly more CCL5 than P2X7R − / − organoids in response to LPS stimulation (Figure 1h; P = 0.046). P2X7R activation is known to induce caspase-1-17 and caspase-11-dependent cell death. 18 As we did not see evidence of epithelial production of the inflammasome-associated cytokines IL-1β and IL-18, we investigated whether the epithelial P2X7R was involved in cell death in T. gondii infection. Thus, we analysed the proportion of apoptotic cells in CMT-93 cells infected with T. gondii for 24 h using flow cytometry. Compared with naive cells, T. gondii infection caused a significant increase in CMT-93 cell death ( Figure 1i). However, P2X7R inhibition did not alter the percentage of dead cells in response to infection (Figure 1i), suggesting that the P2X7R was not involved in T. gondii-induced epithelial cell death.

P2X7R in epithelial cells promotes CCL5 production in vivo
To confirm the role of P2X7R in epithelial cell chemokine production in vivo, WT and P2X7R − / − mice were infected with T. gondii, and chemokine production in intestinal epithelial cells was analysed at days 0 and 1 post-infection (p.i.) by quantitative PCR (qPCR) (Figure 2a). We investigated the expression of CCL5 as we had previously described epithelial cells as a source of this chemokine. Similar to the in vitro data CCL5 was upregulated by infection (Po0.01; Figure 2a). In contrast, CCL5 expression levels at day 1 p.i. remained low in P2X7R − / − epithelial cells (Figure 2a). We next investigated levels of IL-1β and the results were consistent with our in vitro data in that we saw no difference in expression between WT and P2X7R − / − in epithelial cells (data not shown). As the P2X7R has also been associated with production of IL-33, we assayed IL-33 but saw no difference between WT and P2X7R − / − epithelial cells. As there was significantly less early epithelial CCL5 expression in P2X7R − / − mice, the serum chemokine levels of the animals at day 1 p.i. were assessed. Consistent with the qPCR data, P2X7R − / − mice had significantly lower serum CCL5 levels when compared with WT mice (57 ± 9.7 pg ml − 1 in WT vs 26 ± 5.2 pg ml − 1 in P2X7R − / − ) (Figure 2b; Po0.01).
To determine whether the altered chemokine response in T. gondiiinfected P2X7R − / − mice was also true for another small intestinal infection, we also analysed epithelial CCL5 in T. spiralis infection. Similar to the T. gondii data, both WT and P2X7R − / − mice had increased expression of epithelial CCL5 relative to naive mice following infection ( Figure 2c) and CCL5 expression was significantly reduced in P2X7R − / − epithelial cells (Figure 2c; 10 ± 3.3 in WT vs 4.5 ± 4.0 in P2X7R − / − , P = 0.042). Consistent with the qPCR data, P2X7R − / − mice had significantly lower serum CCL5 levels in response to T. spiralis when compared with WT mice (79 ± 21 pg ml − 1 in WT vs 55 ± 8.5 pg ml − 1 in P2X7R − / − ) (Figure 2d, P = 0.03). Collectively, these data confirm our in vitro data that P2X7R deficiency impairs the epithelial chemokine response to infection.
As T. gondii causes damage to the gut and T. spiralis also burrows into intestinal epithelial cells causing damage and intestinal epithelia cells express low levels of P2X7R, 19 we investigated the role of epithelial P2X7R in the apoptotic response to infection. The analysis of ileum sections using TUNEL (terminal deoxynucleotidyl transferase dUTP nick-end labelling) assay showed that P2X7R − / − mice had significantly less cell death in response to infection as compared with WT mice (Supplementary Figure 3). This is in line with published work showing that P2X7R promotes cell death, but as P2X7R is expressed by multiple cell types in the gut, it is not possible to ascribe whether P2X7R expressed in the epithelium is driving cell death. Collectively, however, these data imply a role for epithelial P2X7R in the early response to infection-induced cellular damage.
Impaired recruitment of CD103 + DC subsets to the small intestinal epithelium in infected P2X7R − / − mice We have previously shown that epithelial chemokine production promotes DC recruitment to the gut during infection. 12,15 To assess . Cell supernatants were analysed by ELISA for TNF-α (b), IL-6 (c), CCL5 (d) and IL-18 (e). Small intestinal crypt cells were isolated from C57BL/6 (grey bars) or P2X7R − / − mice (hatched bars), cultured as ex vivo crypt organoids and treated with or without LPS (20 μg ml − 1 ) for 6 h. Cell supernatants were analysed by ELISA for IL-1β (f), IL-18 (g) and CCL5 (h). (i) Following infection by PRU (in a ratio of 1:1) for 24 h, CMT-93 cells with or without the P2X7R antagonist A-740003 treatment (500 μM) were collected for analysis of cell death by Annexin-V/PI (propidium iodide) staining. (a) Apoptotic cells were quantified, indicating that A-740003 treatment had no effect on infection-induced cell death. The data are means ± s.d. pooled from three independent experiments. ND, not detectable. *Po0.05, **Po0.01 and ***Po0.001. Statistical difference was measured using two-way ANOVA with Bonferroni post-test.
There was not a general reduction of DC recruitment into the epithelial layer of the P2X7R − / − intestine. The proportion of CD103 − DCs (CD11c + CD11b + CD103 − F4/80 − ) were unchanged between WT and P2X7R − / − mice in the epithelial layer or LPL compartment after either T. gondii or T. spiralis infection (data not shown). Furthermore, the proportions of F4/80 + CD11c + macrophages were similar between WT and P2X7R − / − before and after infection irrespective of the type of infection (Supplementary Figures 4C,D,G and H). There were no significant differences in the absolute number of CD45 + cells between genotypes at all time-points studied. Collectively, these data show that irrespective of the infection, there is a selective reduction in the early recruitment of DC subsets in the absence of P2X7R.
P2X7R − / − mice have reduced parasite specific T-cell responses and higher parasite burdens p.i. To assess whether the delayed recruitment of CD103 + DCs in response to infection influenced the development of T-cell responses, splenocytes were collected from WT and P2X7R − / − mice at day 8 p.i. and the T-cell response to infection was analysed. For T. gondii infection, splenic CD4 + T cells were assessed for intracellular cytokine production of interferon-γ (IFN-γ) (Figure 4a). There was a significant increase in the proportion of IFN-γ + cells in WT mice compared with P2X7R − / − (Figure 4a; Po0.001). In response to Trichinella infection, we saw a significant reduction of IL-4 + -secreting cells in P2X7R − / − mice compared with WT mice (Po0.001; Figure 4b). To confirm whether the reduced T-cell cytokine productions altered the adaptive The observation that there was a reduced immune response in P2X7R − / − mice suggested that parasite burden and pathology may also be altered. Therefore, parasite burden in the ileum villi was enumerated. Compared with WT mice, P2X7R − / − mice had significantly higher T. gondii tachyzoite burden at day 5 p.i. (Po0.05) (Figures 4e and f). Similarly, P2X7R − / − mice had a significantly higher worm burden (Po0.05) at day 12 p.i. (Figure 4g). These data implicate a role for P2X7R in controlling the local infection of T. gondii and expulsion of T. spiralis.

P2X7R on epithelial cells promotes DC recruitment to the site of infection
P2X7R is expressed by DCs and epithelial cells; therefore, our KO mice lack functional P2X7R in antigen-presenting cells and epithelial cells. To dissect whether the reduction of early CD103 + DC infiltration to the small intestine in P2X7R − / − mice was because of a defect in the capacity of P2X7R − / − DC to migrate in response to chemokines, we generated bone marrow-derived DCs (BMDCs) from WT and P2X7R − / − mice and monitored BMDC migration in response to CCL5 or CCL20, chemokines previously reported to promote DC recruitment to the epithelium. 12,15 Both CCL5 and CCL20 induced a two-to threefold increase of BMDC migration and there was no difference of the migratory response between WT and P2X7R − / − BMDCs (Figures 5a and b).

DISCUSSION
Pattern recognition receptors recognise the highly conserved molecular patterns present on pathogens and also commensal organisms. However, most commensals are not invasive. Therefore, a feature of infection is the damage caused, which is characterised by the release of damage-associated molecular pattern molecules such as ATP. Extracellular ATP is a danger signal that alerts the immune system of abnormal death and threat. Our in vivo data indicated that the P2X7R on epithelial cells has a critical role in sensing infectioninduced damage. Activation of P2X7R in macrophages and DCs has been linked to activation of the inflammasome. The inflammasome serves as a platform for the activation of caspase-1-mediated cleavage and maturation of IL-1β and IL-18. However, in vitro infection or treatment of intestinal epithelial cells by T. gondii or LPS did not trigger an IL-1β or IL-18 response regardless of whether the P2X7R was present, suggesting that in epithelial cells P2X7R may have an alternative role. Our data implicated P2X7R in promoting the production of cytokines and chemokines from intestinal epithelial cells. Among these epithelial-derived C-C chemokines, we have shown that CCL2, CCL5 and CCL20 are potential chemoattractants for intestinal DC homing to the site of infection. 12,15 Our in vivo and in vitro data show that the absence of P2X7R results in impaired CCL5 production. Multiple TLRs are involved in sensing T. gondii infection and initiating proinflammatory responses. TLR11/12 are required for recognising toxoplasma profilin to regulate DC-derived IL-12 in response to T. gondii infection. 21,22 TLR2 and TLR-adaptor MyD88 − / − mice are more susceptible to T. gondii with decreased production of chemokines. 21,23,24 Furthermore, TLRs are known to trigger NF-κB-dependent production of cytokines (e.g., IL-1β, IL-6, TNF-α and type-I interferons) 25 and chemokines, including CXCL10, CCL2 and CCL5, 26 to elicit inflammation in response to infection. Thus, to study whether TLR signalling is involved in the P2X7Rdependent CCL5 response, we used LPS, known to induce CCL5 expression in rodent epithelial cells, to stimulate TLR4 in primary crypt organoids. We found that LPS-induced CCL5 production in primary organoids and that the CCL5 upregulation was reduced in the absence of P2X7R. Our ex vivo data showed that CCL5 production in response to LPS treatment of WT organoids was also lower in P2X7R − / − organoids. These data suggest that blocking P2X7R signalling attenuates TLR activation-induced chemokine production. Our observation that in the absence of the P2X7R, epithelial responses to infection in vivo and in vitro were also reduced suggest that the P2X7R may have a regulatory role in pathogen-pattern recognition receptor-mediated proinflammatory signalling.
CCL5 has been shown to drive the recruitment of several cell types including T cells, macrophages, eosinophils and basophils. 27 Our previous work implicated epithelial chemokine responses in promoting the recruitment of DCs to the site of infection; therefore, we focused on DC recruitment as an indicator of the biological effect of reduced epithelial chemokine signalling. However, it may be that other cells are similarly affected. 12 The reduction of epithelial chemokine production in P2X7R − / − mice was associated with delayed infiltration of CD103 + DCs in response to T. gondii and T. spiralis infection. Early DC responses are known to alter disease outcome to intracellular pathogen infections including T. gondii and Listeria monocytogenes. 28,29 Intestinal CD103 + DCs can be divided into two main subsets, CD103 + CD11b + and CD103 + CD11b − DCs, which differ in terms of distribution, requirement for transcription factors and in vivo function. 30 CD103 + CD11b − DCs are dominant in colon lamina propria and extra-intestinal tissue, 31 whereas CD103 + CD11b + DCs are the major subset in the small intestinal lamina propria and villus. 30 Additionally, the CD103 + CD11b − (CD8α + ) DC subset in the small intestine expresses several different TLRs and promotes Th1 immunity and CTL activity. 32 In the T. gondii-infected mouse ileitis model, we noted that blocking P2X7R delayed the recruitment of CD103 + CD11b − DCs, whereas CD103 CD11b + DCs were reduced in T. spiralis infection. CD103 + CD11b − DCs are thought to be the major IL-12-producing immune cell against acute T. gondii infection 33 and have an important role in cross-presentation. 34 Thus, the intraepithelial recruitment of CD103 + CD11b − DCs observed in WT mice could be important for early control of T. gondii infection. Indeed, our observations that there was a higher parasite load in P2X7R mice could corroborate this, although the altered recruitment of other innate cells such as neutrophils may contribute to this also. The in vivo role of CD103 + CD11b + DCs in small intestinal inflammation is not yet clear, but they are thought to be critical in both tolerogenic and inflammatory responses, 33,35 and selective reduction of CD103 + CD11b + DCs by depleting interferon regulatory factor 4 has been shown to impair Th17 differentiation in mesenteric lymph node. 36 Thus, it is still unclear which CD103 + DC subset has a more important role against Th1 or Th2 infections. A previous report showed that defective chemoattraction of small intestinal CD103 + DCs in neonatal mice impairs the development of protective immunity against Cryptosporidium parvum infection. 37 Thus, the delayed chemoattraction of CD103 + DC subsets may delay the infiltration of CD103 + DCs into the draining lymph nodes for T-cell priming and correlate with the impaired Th immunity seen in P2X7R − / − mice.
Collectively, our findings identify a novel role for small intestinal epithelial P2X7R in the induction of epithelia cytokine responses and early DC infiltration in response to infection. Given the proinflammatory role as a potential tissue damage sensor to initiate inflammation, P2X7R is likely to be an important target for the development of new therapies for inflammatory disorders in the gut.

METHODS Mice
C57BL/6 (WT) and P2X7R − / − littermate controls were generated from breeding heterozygous Pfizer P2X7R − / − mice (The Jackson Lab, Bar Harbor, ME, USA). 20 All mice were maintained by the Biological Services Facility (University of Manchester, UK), and kept in individually ventilated cages and fed a standard chow. Experiments were performed in accordance with the Home Office Animals (Scientific Procedures) Act (1986). Three to six mice per group were used per study and each infection was repeated two to three times. For T. gondii infection, age-matched male mice each mouse was orally infected with 1 × 10 6 of tachyzoites in 0.1 ml phosphate-buffered saline (PBS) in the morning. The maintenance, recovery and infection of T. spiralis was described previously. 38 Age-matched male mice were infected with 400 larvae by oral gavage. Worm burden was assessed by longitudinal section of the small intestine followed by incubation in PBS at 37°C for 4 h. Mice were monitored throughout infection and no unexpected adverse effects were observed.

T. gondii and cells
T. gondii type-I RH strain expressing yellow fluorescence protein 39 and type-II Prugniaud (PRU) expressing tandem dimers of tomato-red fluorescent protein T. gondii strain was used in our experiments. 40 The tachyzoites were harvested by serial 4-5 days passage in human foreskin fibroblast cultured in Dulbecco's modified Eagle's medium (Sigma-Aldrich, Dorset, UK) with 10% fetal bovine serum (FBS) (Sigma-Aldrich) and 1% penicillin/streptomycin (Sigma-Aldrich). The mouse colonic epithelial cell line, CMT-93, was maintained in Dulbecco's modified Eagle's medium (Sigma-Aldrich) with 10% FBS (Sigma-Aldrich), 2 mM L-glutamine (Sigma-Aldrich) and 1% penicillin/streptomycin (Sigma-Aldrich). CMT-93 cells were infected with PRU at a ratio of 1:5 for 24 h and treated with the selective P2X7R antagonist, A-740003 (500 μM; Sigma-Aldrich) to block P2X7R at the time of infection. A total of 1 × 10 6 cells per group cells were infected by T. gondii PRU strain in a ratio of 1:1. THP-1 monocyte cells were primed using 0.1 μg ml − 1 of LPS for 24 h followed by infection with T. gondii for 24 h in a ratio of 1:4.

Isolation of small intestinal cells
Cells from the small intestine were isolated as described previously. 41 Briefly, Peyer's patches were removed, and the remaining tissue cut into pieces and transferred into Hanks' balanced salt solution supplemented with 1% HEPES and 2 mM ethylenediaminetetraacetic acid. The tissue pieces were incubated at 37°C with shaking for three sets of 15 min and the cells collected and pooled. Intestinal epithelial cells were isolated by Percoll density gradient isolation (Scientific Laboratories Supplies, Yorkshire, UK) and resuspended in RPMI-1640 (Sigma-Aldrich) with 10% FBS and 1% HEPES. The remaining tissue fragments were incubated in RPMI media supplemented with collagenase VIII and CaCl 2 at 37°C for 60 min to collect LPL. All the epithelial layer cells and LPL were resuspended at 1 × 10 7 cells per ml in FACS buffer (PBS plus 2% FBS with 0.1% sodium azide) for flow cytometry analysis.

ELISA and cytokine analysis
IL-1β, IL-18 and CCL5 levels in mouse serum and supernatants from CMT-93 and crypt organoids were measured using ELISA DuoSet Kits (R&D Systems) according to the manufacturer's instructions. ATP was measured using the Luciferin-Luciferase Bioluminescence Assay (Thermo Fisher Scientific, Loughborough, UK) as per the manufacturer's instructions on three technical replicates. Levels of IL-4, IL-10, IL-6, IL-9, IL-13, IFN-γ, TNF-α and IL-12p70 in the serum were determined using a cytometric bead array according to the manufacturer's instructions (BD) and analysed using BD FacsAria cytometer (BD) and FCAP Array software (BD).