Extracellular ATP Activates the NLRP3 Inflammasome and Is an Early Danger Signal of Skin Allograft Rejection.

Immune cells are equipped with a number of receptors that recognize sterile injury and pathogens. We find that host immune cells release ATP as an inflammatory signal in response to allogeneic transplantation. ATP then acts via a feedback mechanism on the P2X7 channel to activate the NLRP3 inflammasome and subsequently process and release interleukin (IL)-18. This process is a necessary stage in the deleterious Th1 response against allotransplantation via interferon-γ production. Lack of IL-18 resulted in a decrease in graft-infiltrating CD8 cells but an increase in regulatory T cells. In human liver transplant patients undergoing progressive immunosuppressive drug withdrawal, we found that patients experiencing acute rejection had higher levels of the P2X7 receptor in circulating inflammatory monocytes compared to tolerant patients. These data suggest that the pharmacological inhibition of the P2X7 receptor or the NLRP3 inflammasome will aid in inducing transplant tolerance without complete immunoparalysis.

Introduction

Transplantation of antigen-containing tissues between non-genetically identical individuals is referred to as allogeneic organ transplantation. This procedure is growing in popularity, with over 38,000 procedures occurring in 2016 in the United States alone (Bentley and Phillips, 2017). However, in allogeneic organ transplantation, antigens in graft tissues can lead to non-self-recognition and, thus, generate a destructive immune response in the form of tissue rejection. Nevertheless, non-infected transplants are sterile and are beneficial rather than dangerous to the recipient organism (Land, 2012). The allogeneic immune response is initiated when innate antigen-presenting cells present non-self-antigens to lymphocytes, resulting in a destructive adaptive Th1 response to grafts. Tissue damage caused by the transplantation procedure results in the release of damage-associated molecular patterns (DAMPs), which activate innate immunity during allograft rejection. These DAMPs include the induction of heat shock proteins, the production of reactive oxygen species, hypoxia, and the degradation of the extracellular matrix (Critchley and Fildes, 2012, Land, 2012, Matzinger, 2002). Recently it has also been demonstrated that innate antigen-presenting cells recognize and are directly activated by alloantigens (Oberbarnscheidt et al., 2014, Zecher et al., 2009). However, the way that endogenous DAMPs are regulated and elicit an allogeneic immune response is poorly understood.

The pro-inflammatory cytokine interleukin (IL)-1β plays a key role in host responses to infection and tissue injury. IL-1β is expressed in an inactive pro-form, which is cleaved into an active form by the protease caspase-1 (Cassel and Sutterwala, 2010). Activation of caspase-1 is regulated by a group of large multi-protein complexes referred to as inflammasomes formed by cytosolic pattern recognition receptors (PRRs), often of the nucleotide-binding domain- and leucine-rich repeat-containing receptor (NLR) family. The best-characterized inflammasome is NLRP3 (Cassel and Sutterwala, 2010). Inflammasomes are composed of the PRR and pro-caspase-1 and, in some types, the adaptor apoptosis-associated speck-like protein containing a caspase recruitment domain (ASC) that interacts via homology, thus binding domains with the PRR and pro-caspase-1 (Latz et al., 2013). Activating the NLRP3 inflammasome requires a priming signal to increase the expression of the receptor and of pro-IL-1β and a second signal (i.e., the activation of the P2X7 receptor by extracellular ATP) to induce inflammasome formation, activate caspase-1, and process pro-IL-1β into a mature cytokine that is secreted from the cell via unconventional protein release mechanisms (Cassel and Sutterwala, 2010, Latz et al., 2013).

Extracellular ATP is a key DAMP that induces NLRP3 inflammasome formation by signaling through the purinergic P2X7 cell surface receptor (Di Virgilio et al., 2017, Ferrari et al., 2006, Idzko et al., 2014). The P2X7 receptor is associated with two different membrane permeabilities: (1) a fast cation conductance that increases cytosolic Ca2+ and decreases intracellular K+, these being key triggers of NLRP3 inflammasome activation; and (2) a second permeability after prolonged ATP stimulation that allows the passage of molecules of up to ∼900 Da in molecular weight (Pelegrín, 2011a). NLRP3 and the subsequent caspase-1 activation also regulate the release of IL-18, another pro-inflammatory cytokine that is a potent inducer of the Th1 response because it produces interferon (IFN)-γ (Novick et al., 2013). The IL-18 precursor is constitutively expressed in cells and does not require gene induction (Novick et al., 2013); this difference in expression when compared to IL-1β is critical in conditions where there is no priming signal and, therefore, no pro-IL-1β. A clear example of this can be seen in the autoinflammatory syndromes caused by mutations in NLRP3, where systemically elevated levels of IL-18 are present but IL-1β is not usually detectable (Baroja-Mazo et al., 2014, Novick et al., 2013). Therefore, IL-1β and IL-18 will affect pathogenesis differently depending on the nature of the disease or injury. Allotransplantation rejection involves a substantial innate and adaptive immune response in which the roles of IL-1β and IL-18; their upstream regulators NLRP3, caspase-1, and extracellular ATP; and the P2X7 receptor have yet to be fully understood.

In this study, we found that the innate immune system is activated by extracellular ATP during allotransplantation. We observed that soon after allogeneic (but not syngeneic) transplantation, antigen-presenting cells modulated extracellular ATP independently of cell death and via a mechanism involving the P2X7 receptor. This signaling activates the NLRP3 inflammasome and releases IL-18 to ensure allograft rejection. We demonstrate that pharmacological blockage of the P2X7 receptor decreases the concentration of extracellular ATP and increases graft survival and that this could represent a new procedure for developing graft tolerance without complete immunosuppression. Our work also raises clinical awareness of the need to correctly manage DAMPs in transplantation medicine.

Results

Increased Extracellular ATP in Allogeneic Transplantation

Using the pmeLuc biosensor (Pellegatti et al., 2005), we found that at 3 days after conducting separate allogeneic and syngeneic skin transplants in mice, extracellular ATP was higher in the allogeneic transplants and was locally associated with the grafted allogeneic skin (Figures 1A and 1B). We confirmed that, in the control group at 14 days post-transplantation, allogeneic tissue was rejected whereas syngeneic tissue was tolerated (Figure S1A). Degradation of extracellular ATP by local injection of exogenous apyrase reduced the bioluminescence signal in allografts, but it did not significantly decrease the bioluminescence signal from an intracellular luciferase reporter assay (Figures S1B and S1C), thus confirming that the bioluminescence signal of pmeLuc was due to extracellular ATP.

Figure 1

Allogeneic Transplants Are Associated with Extracellular ATP

(A) Representative images of extracellular ATP 3 days after syngeneic or allogeneic skin transplantation, using C57BL/6 as recipient and C57BL/6 (syngeneic) or BALB/c (allogeneic) skin as donor.

(B) Quantification of graft-associated bioluminescence as the average radiance (p/s/cm2/sr) at day 3 after transplantation (n = 8).

(C) Extracellular ATP quantification at 3, 7, and 14 days after syngeneic or allogeneic transplantation (n = 7−10).

(D) Representative images of H&E-stained grafts. Arrows denote tissue necrosis and the detachment of epidermal layer. Scale bar, 50 μm.

(E) CD3+ cells 3 or 7 days after transplantation (n = 12−18 fields of view from 3–4 transplants).

(F) Relative gene expression for the indicated genes in grafts 3 or 7 days after transplantation (n = 4–5).

(G) IFN-γ ELISA produced by axillar lymph node lymphocytes collected 7 days after syngeneic or allogeneic transplantation and subjected to mixed lymphocytic reaction with allogeneic splenocytes (n = 3–6).

Data are represented as mean ± SEM (B, C, and E–G). Student’s t test was used for (B) and ANOVA with Bonferroni post-test was used for (E)–(G). See also Figure S1.

Extracellular ATP is considered a tissue damage-associated molecule (Idzko et al., 2014), and it could be passively released from cells during the process of allogeneic tissue destruction (Oberbarnscheidt and Lakkis, 2014). However, we found that, while extracellular ATP decreased over time after allogeneic tissue transplantation (Figure 1C), allogeneic tissue damage actually increased (Figure 1D). Therefore, elevated extracellular ATP in the early stages of allotransplantation cannot be considered a direct consequence of surgery, the allogeneic immune response, or the associated tissue damage. Despite elevated local extracellular ATP 3 days after allotransplantation (Figure 1C), the allogeneic-associated adaptive immune response was not observed until day 7, when we found an increase in infiltrating CD3-lymphocytes and a robust production of the IFN-γ and IL-12 cytokines (Figures 1E–1G; Figure S1D). This late adaptive immune response is expected in skin transplants that are not immediately vascularized. This is in contrast to rapidly vascularized grafts where T cells infiltrate the graft within hours and produce IFN-γ (Su et al., 2014). The slower adaptive response in the skin model suggests that extracellular ATP is produced as an early inflammatory signal in allografts before the Th1 allogeneic immune response is generated.

The P2X7 Receptor Modulates ATP Release in Allografts

Extracellular ATP is a danger signal that triggers macrophage NLRP3 inflammasome by activating the purinergic P2X7 receptor (Idzko et al., 2014, Pelegrin, 2011b, Schroder and Tschopp, 2010). Consequently, we used a genetic approach to study the ATP release mechanism in allografts. We found that ATP release was absent in P2xr7−/− mice 3 days after allotransplantation (Figure 2A), suggesting that the P2X7 receptor modulates the extracellular levels of its own agonist in vivo. This was also confirmed when wild-type mice were treated with the selective P2X7 receptor antagonist A438079 (Donnelly-Roberts and Jarvis, 2007); the wild-type mice showed a significantly lower extracellular ATP response to allogeneic grafts than vehicle-treated mice (Figure 2B).

Figure 2

Allografts Induce ATP Release via P2X7 Receptor in Antigen-Presenting Cells

(A) Extracellular ATP associated with allografts 3 days after transplantation, using C57BL/6 wild-type or P2rx7−/− as recipient and BALB/c skin as donor. Left: representative mouse per group is shown. Right: average radiance (p/s/cm2/sr) of n = 5−8 transplants is shown.

(B) Extracellular ATP associated with allografts 3 days after transplantation in wild-type mice treated with vehicle or A438079 (100 μmol/kg). Left: representative mouse per group is shown. Right: average radiance (p/s/cm2/sr) of n = 5−6 transplants is shown.

(C and D) ATP release from C57BL/6 bone marrow-derived macrophages (BMDMs), (C) incubated for 2 hr with C57BL/6 or BALB/c skin (n = 4) or (D) incubated for 30 min in the absence or presence of ovalbumin (OVA, 10 mg/mL) and A438079 (25 μM) (n = 3).

(E) Yo-Pro uptake kinetic in BMDMs treated or not for 2 hr with OVA (10 mg/mL) or LPS (1 μg/mL) and then stimulated as indicated by an arrow with ATP (3 mM) in the presence or absence of A438079 (25 μM) (n = 6).

(F) Extracellular ATP associated to allografts 3 days after transplantation in P2rx7−/− mice treated with 106 wild-type (P2rx7+/+) or P2rx7−/− BMDMs applied under the skin graft during the surgical procedure.

Data are represented as mean ± SEM. For (A), (B), and (F), left shows representative mouse per group, and right shows p/s/cm2/sr of n = 5−6 transplants. Student’s t test was used for (A), (B), and (F); ANOVA with Bonferroni post-test was used for (C) and (D) and with Tukey post-test for the slopes represented in (E). See also Figure S2.

We next found that the P2X7 receptor was mainly expressed to roughly the same levels by macrophages in both syngeneic and allogeneic grafts (Figures S2A–S2C). The decrease in extracellular ATP found at 3 days post-transplantation in P2xr7−/− mice was not due to a decrease in the amount of graft-infiltrating macrophages (Figure S2D). Extracellular ATP increased in supernatants from macrophages incubated in vitro with allogeneic, but not syngeneic, skin tissue (Figure 2C) and after incubation with the classic antigen ovalbumin (Figure 2D). Ovalbumin was able to induce ATP release via the P2X7 receptor (Figure 2D) and increase early macrophage plasma membrane permeabilization after activation of the P2X7 receptor (Figure 2E). This suggests that the P2X7 receptor-associated pore could be a pathway for ATP release during antigen processing by macrophages (Pelegrín, 2011a).

We next investigated the role of the macrophage P2X7 receptor in the release of ATP during allogeneic transplantation by supplementing allografts on P2xr7 mice with wild-type or P2xr7 macrophages. In vivo supplementation of wild-type macrophages in the P2xr7−/− host restored local extracellular ATP to about the same level seen in wild-type mice (Figure 2F). Conversely, allogeneic transplantation of P2X7-deficient skin onto wild-type recipients presented a similar level of extracellular ATP when compared to wild-type skin (Figure S2E). This suggests that allograft recognition in vivo by recipient macrophages amplifies the release of ATP via the P2X7 receptor and that donor immune cells do not release extracellular ATP in the skin transplantation model.

Blocking the P2X7 Receptor Reduces Allograft Rejection

We next found lower levels of epithelial hyperplasia in the transplanted skin of P2xr7−/− mice after 3 days (Figures 3A and 3B). Epithelial hyperplasia is an early sign of skin distress (Lee et al., 2009), and we observed low levels of hyperplasia in syngeneic grafted mice compared to the extensive levels in allogeneic grafted mice (Figures 3A and 3B; Figures S3A and S3B). Furthermore, treatment of wild-type allotransplanted animals with A438079 also resulted in decreased epithelial thickness (Figure 3B; Figure S3A), thus demonstrating the therapeutic potential of reducing extracellular ATP by early targeting of the P2X7 receptor during the period immediately after the transplant procedure. Furthermore, in terms of the Banff pathological classification for skin allografts (Cendales et al., 2008), rejection severity was lower at 7 days after transplantation in P2xr7−/− mice (Banff score 3) compared to wild-type animals (Banff score 4). P2xr7−/− mice or mice treated with A438079 also presented decreased inflammatory infiltrate, reduced necrosis, and preserved epidermal and dermal allograft structure at 7 days after transplantation (Figure 3A; Figure S3A).

Figure 3

P2X7 Receptor Blockage Reduces Allograft Rejection

(A) Representative images of H&E-stained allogeneic skin grafts using C57BL/6 wild-type or P2rx7−/− as recipient and BALB/c skin as donor. Scale bar, 50 μm.

(B) Epidermal thickness at 3 days after transplantation. A438079 was intraperitoneally (i.p.) injected (100 μmol/kg/day) (n = 19−71 measurements from 3–4 transplants).

(C) CD3+ or F4/80+cells 7 days after transplantation (n = 6−26 fields of view from 3–4 transplants).

(D) IFN-γ in serum 7 days after transplantation (n = 3–5).

(E) IFN-γ ELISA produced by axillar lymph node lymphocytes collected 7 days after transplantation and subjected to mixed lymphocytic reaction (MLR) with allogeneic splenocytes (n = 3–7).

(F and G) Measurement of transplanted skin by (F) ultrasound or (G) skin transplant survival (n = 6−11 transplants).

(H) Activity of plasma liver enzymes in human serum and mean fluorescence intensity for P2X7 receptor surface expression in inflammatory monocytes (CD14+CD16high) from liver-transplanted patients (tolerant or non-tolerant) during immunosuppression and after immunosuppression withdrawal (rejection) (n = 4 patients/group).

(I) Correlation between the fold change of plasma ATP concentration and liver enzymes during immunosuppression and after immunosuppression withdrawal in the 4 non-tolerant patients.

For (B), (F), and (H), data are represented as mean ± SEM. One-way ANOVA with Bonferroni post-test was used for (B)–(E), two-way ANOVA with Bonferroni post-test was used for (F) and (H) (∗differences between allogeneic and isogeneic; §differences between allogeneic and allogeneic + A438079; ns, no significant difference between isogeneic and allogeneic + A438079), and Gehan-Breslow-Wilcoxon test was used for (G). See also Figures S3 and S4 and Table S1.

P2xr7−/− mice presented reduced levels of specific allogeneic immunoglobulin Gs (IgGs) in serum at 14 days after transplantation (Figure S3C), suggesting that the P2X7 receptor is important in establishing a full allogeneic response driven by B cells. This was paralleled by a decrease in allograft-infiltrating CD3+ T cells and F4/80+ macrophages at 7 days after transplantation in P2xr7−/− mice (Figure 3C; Figure S3D). In addition, allotransplanted mice treated with A438079 showed lower IFN-γ concentrations in serum when compared to allotransplanted wild-type mice (Figure 3D). This was matched by a reduction in IFN-γ production after the expansion of T cells in response to BALB/c alloantigens (Figure 3E), thus suggesting that blocking the P2X7 receptor’s activation and extracellular ATP causes defective T cell activation in allotransplants. This reduced alloimmune response increased the survival rates of allogeneic transplanted tissues (Figures 3F and 3G; Figure S3E). Overall, we have shown that transplantation of allogeneic, but not syngeneic, tissue induces an early release of ATP from innate immune cells via the P2X7 receptor and thus establishes an efficient alloimmune response driven by the T and B cells.

P2X7 Receptor Expression Increases during Human Liver Allograft Rejection

To obtain evidence of the potential importance of the P2X7 receptor during human transplant rejection, we analyzed the expression of the P2X7 receptor on the surface of inflammatory blood monocytes (CD14+CD16high) in liver transplant patients after controlled immunosuppression withdrawal. Some patients rejected the transplant during immunosuppression withdrawal (non-tolerant patients), while other patients became tolerant to it (tolerant patients) (Table S1). After immunosuppression withdrawal, plasma concentration of liver enzymes increased during acute rejection in non-tolerant patients when compared to tolerant patients (Figure 3H; Figure S4A). Liver biopsy corroborated immune rejection of the liver in patients with increased levels of liver enzymes in plasma (data not shown). During rejection, P2X7 receptor expression in circulating inflammatory monocytes increased (Figure 3H) without changing the percentage of inflammatory monocytes (Figure S4B). In tolerant patients, P2X7 receptor expression was similar both during immunosuppression and after withdrawal (Figure 3H). The fold change of plasma ATP concentration measured during immunosuppression and after rejection in non-tolerant patients positively correlated with the fold change in plasma liver enzymes (Figure 3I), suggesting that extracellular ATP and the P2X7 receptor could play a role in the immune allogeneic response during rejection of liver transplants in humans.

The P2X7 Receptor Engages the NLRP3 Inflammasome in Allografts

We next analyzed whether P2X7 receptor-induced pyroptosis could be the mechanism for ATP release in allografts, given that the P2X7 receptor is a well-known activator of caspase-1 via the canonical NLRP3 inflammasome pathway and that caspase-1 is involved in pyroptosis, a specific type of cell death characterized by the release of intracellular components (Miao et al., 2011). However, we did not find any difference in extracellular ATP levels in allografts from wild-type versus Nlrp3−/− or Casp1/11−/− mice (Figures 4A and 4B). These data suggest that, upon alloantigen recognition, pyroptosis mediated by activation of either caspase-1 or caspase-11 does not induce ATP release but rather the P2X7 receptor in macrophages induces a pathway for ATP release (Pellegatti et al., 2005), which in turn could feed back into the P2X7 receptor itself.

Figure 4

The NLRP3 Inflammasome Is Not Involved in ATP Release in Allografts but Is Activated in Allografts

(A) Extracellular ATP associated with allografts 3 days after transplantation in C57BL/6 wild-type (Nlrp3+/+) or Nlrp3−/− mice as recipients and BALB/c skin as donor. Left: luciferase signal representing extracellular ATP in one representative mouse per group is shown. Right: average radiance (p/s/cm2/sr) of n = 7−8 transplants is shown.

(B) Extracellular ATP associated with allografts 3 days after transplantation in C57BL/6 wild-type (Casp1/11+/+) or Casp1/11−/− mice as recipients and BALB/c skin as donor. Left: luciferase signal representing extracellular ATP in one representative mouse per group is shown. Right: average radiance (p/s/cm2/sr) of n = 3−4 transplants is shown.

(C) Relative gene expression for the indicated genes in grafts 3 days after transplantation, using C57BL/6 as recipient and C57BL/6 (syngeneic) or BALB/c (allogeneic) skin as donor. A438079 was i.p. injected (100 μmol/kg/day) (n = 5−6). Gene expression of healthy skin is presented with a dashed line.

(D) IL-18 ELISA in serum 3 days after transplantation (n = 3−5).

Data are represented as mean ± SEM. ns, no significant difference (p > 0.05); Student’s t test was used for (A) and (B); ANOVA with Bonferroni post-test was used for (C) and (D).

It is well known that the P2X7 receptor activates the NLRP3 inflammasome, this being a key event in the activation of the adaptive immune system (Pelegrin, 2011b, Schroder and Tschopp, 2010). We found that Casp1 and Il1b gene expression was upregulated in allografts at 3 days post-transplantation and that this upregulation was reduced by A438079 treatment (Figure 4C). Nlrp3 gene expression was also upregulated at 3 days post-transplantation to the same extent in both syngeneic and allogeneic transplants (Figure 4C). This suggests that Nlrp3 expression might be caused by tissue damage during surgery, as has been demonstrated in a sterile injury model (Komada et al., 2015). We also found constitutive expression of the Il18 gene in both syngeneic and allogeneic grafts (Figure 4C), in line with the fact that IL-18 is expressed as a constitutive gene with its release controlled by the activation of caspase-1 (Novick et al., 2013). Accordingly, the circulating blood concentration of IL-18 increased in allotransplanted mice after 3 days and reduced in allotransplanted mice treated with A438079 (Figure 4D). Despite increased local Il1b gene expression in the allografts (Figure 4C), this cytokine was not identified at a systemic level in the serum at days 3 or 7 post-transplantation. This indicates that the P2X7 receptor could regulate the NLRP3 inflammasome at an early stage after allotransplantation to induce systemic IL-18 cytokine in vivo.

NLRP3 Inflammasome Signaling Contributes to the Severity of Allograft Rejection

Several inflammasome-related genes were further upregulated in allografts at day 7 post-transplantation (Figures 5A–5C). Upregulation of Casp1 gene expression in allografts was reduced in Nlrp3−/− mice (Figure 5A), while Nlrp3 and Il1b upregulation was significantly decreased in Casp1/11−/− mice (Figures 5B and 5C). These data show that early induction of the inflammasome by P2X7 receptor signaling is persistent (up to at least 7 days after allogeneic transplantation) via a positive feedback loop induced by the NLRP3 inflammasome itself.

Figure 5

Inflammasome Is Upregulated during Graft Rejection

(A) Casp1 gene expression in grafts (n = 3−6).

(B and C) Relative gene expression for (B) Nlrp3, (C) Il1b, and Il18 in grafts (n = 3–6).

Data are represented as mean ± SEM. ANOVA with Bonferroni post-test was used for (A)–(C), using C57BL/6 wild-type, Nlrp3−/−, or Casp1/11−/− as recipient and C57BL/6 (syngeneic) or BALB/c (allogeneic) skin as donor. WT, wild-type.

We next found that NLRP3 inflammasome deficiency reduced epithelial hyperplasia after 3 days (Figures 6A and 6B; Figure S5A). Histological tissue damage observed 7 days after transplantation was less severe in allotransplanted Nlrp3−/− and Casp1/11−/− mice than in allotransplanted wild-type mice (Figure 6A; Figure S5B). Furthermore, when compared to wild-type mice, Nlrp3−/− and Casp1/11−/− allotransplanted mice presented significantly reduced circulating IL-18 cytokine (Figure 6C) and a decreased number of allograft-infiltrating CD3+ T cells and F4/80+ macrophages (Figure 6D; Figures S5C and S5D). In addition, 7 days after transplantation, allotransplanted Nlrp3−/− and Casp1/11−/− mice showed low Ifng gene expression in the allograft and lower IFN-γ serum concentration when compared to allotransplanted wild-type mice (Figures 6E and 6F). This was paralleled by reduced IFN-γ production when T cells expanded in response to BALB/c alloantigens in mixed leukocyte reactions (Figure 6G), thus suggesting that a lack of NLRP3 inflammasome causes defective T cell activation in allotransplants.

Figure 6

NLRP3 Inflammasome Deficiency Reduces Allogeneic Immune Response

(A) Representative images of H&E-stained skin grafts, using C57BL/6 wild-type, Nlrp3−/−, or Casp1/11−/− as recipient and BALB/c skin as donor. Scale bar, 50 μm.

(B) Epidermal thickness at 3 days after transplantation (n = 20−70 measurements from 3–4 transplants).

(C) IL-18 ELISA in serum at 3 days post-transplantation (n = 4−6).

(D) CD3+ or F4/80+cells 7 days after transplantation (n = 6−20 fields of view from 3–4 transplants).

(E and F) IFN-γ (E) relative gene expression or (F) serum concentration 7 days after transplantation (n = 3−4, E; n = 3–8, F).

(G) IFN-γ ELISA produced by axillar lymph node lymphocytes collected 7 days after transplantation and subjected to mixed lymphocytic reaction (MLR) with allogeneic splenocytes (n = 4–7).

For (F) and (G), each circle represents an individual mouse. For (B)–(G), data are represented as mean ± SEM. ANOVA with Bonferroni post-test was used for (B)–(G). See also Figure S5.

IL-18 Mediates Allograft Rejection

To determine whether the increase in IL-18 could be mediating the allograft rejection response, we first measured the IL-18-binding protein (IL-18BP) in the serum of transplanted animals to verify that enough free IL-18 was available for signaling. Allogeneic transplantation did not increase IL-18BP concentration, and the levels of free IL-18 were significantly higher for allogeneic transplantation when compared to syngeneic transplantation (Figure 7A). We found that IL-18 was not detectable in the serum of IL-18 knockout mice (Figure S6A) and that their concentrations of IL-18BP were not affected (Figure S6A). However, the amounts of IL-12p40 and IFN-γ were significantly reduced in IL-18-deficient mice when compared to wild-type animals (Figures 7B and 7C). Furthermore, a mixed lymphocyte reaction produced less IFN-γ when the lymphocytes derived from the allogeneic transplanted IL-18 knockout mice (Figure 7D). The amount of IL-12p40 and IFN-γ was not reduced in IL-1 receptor 1-deficient mice (Figures S6B and S6C), suggesting that IL-18 (and not IL-1β) plays a role in the inflammasome-mediated allogeneic response.

Figure 7

IL-18 Mediates Inflammasome Alloimmune Response

(A) IL-18-binding protein (IL-18BP) (left) and the ratio of IL-18/IL-18BP (right) in serum 7 days after transplantation (n = 4–5) using C57BL/6 as recipient and C57BL/6 (syngeneic) or BALB/c (allogeneic) skin as donor.

(B) Relative gene expression for Il12b in grafts 3 or 7 days after transplantation using C57BL/6 wild-type or Il18−/− as recipient and BALB/c skin as donor (n = 3–4).

(C) IFN-γ in serum 7 days after transplantation (n = 3–4).

(D) IFN-γ ELISA produced by axillar lymph node lymphocytes collected 7 days after transplantation and subjected to mixed lymphocytic reaction (MLR) with allogeneic splenocytes (n = 4).

(E) Representative images of H&E-stained skin grafts, using C57BL/6 wild-type or Il18−/− as recipient and BALB/c skin as donor. Scale bar, 50 μm.

(F) F4/80+, CD3+, CD8+, or FOXP3+ cells 7 days after transplantation (n = 24–51 fields of view from 3 transplants).

For (C) and (D), each circle represents an individual mouse. For (A)–(D) and (F), data are represented as mean ± SEM. Student’s t test was used for (A)–(D) and (F). See also Figure S6.

This was further confirmed by histology 7 days after allotransplantation. IL-18-deficient mice showed a conserved dermis and cutaneous gland structure (Figure 7E), with a calculated Banff score of 2 compared to a score of 4 in wild-type mice. However, IL-1R1 deficiency resulted in destruction and necrosis of the dermis and epidermis (Figure S6D). Although macrophages in the allogeneic transplant were lower in IL-18-deficient mice than in wild-type mice, CD3+ lymphocyte cell infiltration was similar in both types of mice (Figure 7F). However, the number of CD8+ lymphocytes was significantly lower in IL-18 knockout mice than in wild-type or IL-1R1-deficient mice (Figure 7F; Figures S6E and S6F), and IL-18-deficient mice showed an increase in T-regulatory cells (Figure 7F; Figure S6G). Overall, we show that transplantation of allogeneic, but not syngeneic, tissue induces an early release of ATP from innate immune cells via the P2X7 receptor, which engages the NLRP3 inflammasome and establishes a detrimental alloimmune response via IL-18 signaling.

Discussion

Organ transplantation is a sterile situation where a destructive immune response is developed due to non-self-antigen recognition. It is not clear how DAMPs, particularly extracellular ATP, are regulated to modulate allograft immunity. Here we found that, following skin allotransplants, the P2X7 receptor causes ATP release from innate immune cells and that this signaling activates the NLRP3 inflammasome to release IL-18, which in turn mediates the induction of IFN-γ and the destructive Th1 response to allografts.

During organ transplant surgery or after the killing of grafted cells, there is a release of several DAMPs that activate immunity in allografts (Kono and Rock, 2008, Matzinger, 2002, Oberbarnscheidt and Lakkis, 2014). In particular, extracellular ATP has been found to increase in mouse models of graft versus host disease after tissue damage induced by external radiation to deplete the bone marrow compartment (Wilhelm et al., 2010). In the skin transplantation model used in the present study, syngeneic and allogeneic grafts followed the same surgical procedure, but they resulted in different concentrations of extracellular ATP. During allograft rejection there was a reduction in extracellular ATP, suggesting that the release of ATP in allogeneic transplants was not solely a passive consequence of tissue damage during the surgery procedure or due to immune rejection of grafted cells.

ATP can be released via different active processes independently of cell death (Vénéreau et al., 2015), and here we demonstrate that macrophages release ATP by a mechanism dependent on the P2X7 receptor, thus reiterating the finding of an earlier recombinant cell culture study that demonstrated the role of P2X7 receptor activation in inducing ATP efflux (Pellegatti et al., 2005). Although the P2X7 channel pore is of sufficient size (1.4-nm diameter) to allow the release of ATP (∼0.7 nm) (Browne et al., 2013), the electrostatic cationic profile inside the P2X7 channel could impair the direct transport of the negatively charged ATP molecule. Therefore, the P2X7 receptor may need to combine with accessory anionic channels, such as pannexin-1, to mediate the release of ATP from macrophages during allogeneic transplantation (Ma et al., 2009, Pelegrín, 2011a, Pelegrin and Surprenant, 2006, Schachter et al., 2008). However, we cannot rule out that extracellular ATP in allotransplants could also be a result of necrosis or necroptotic cell death given that the P2X7 receptor has been implicated in necrosis (Dagvadorj et al., 2015). However, in the present study, ATP release was not correlated with tissue damage, suggesting necrosis is not responsible for the elevated extracellular ATP. Recent reports have demonstrated that early innate immune processes in allografts control later adaptive rejection responses by inducing T and B cell migration and activation (Oberbarnscheidt et al., 2014). Our study suggests that extracellular ATP released by macrophages in skin allotransplants is an essential signaling system that drives the events inherent to allotransplantation and the subsequent deleterious adaptive immune rejection. However, we cannot rule out that, in other transplant scenarios, extracellular ATP in allografts could also be due to other tissue-resident cells, such as myocytes or pancreatic islet cells, which have also been reported to release ATP in different pathological scenarios (Clarke et al., 2009, Kowal et al., 2015).

After we had established the role of P2X7 activation and ATP release in allogeneic transplant rejection, we determined the key downstream signaling mechanisms involved, and we identified NLRP3 activation and IL-18 signaling as the principal allogeneic mediators (Idzko et al., 2014). IL-18 is an inflammasome-dependent cytokine that potently induces a Th1 response and, unlike IL-1β, pro-IL-18 is usually expressed basally (Novick et al., 2013); therefore, the single step of inflammasome activation results in the release of mature IL-18. Accordingly, our study found increases in serum IL-18 in response to allograft transplantation but no changes in IL-1β.

We found that NLRP3 is the essential PRR involved in inflammasome formation in allogeneic transplant rejection and that NLRP3-dependent IL-18 maturation, not IL-1β, was necessary to stimulate T cell proliferation, IFN-γ release, and the adaptive allogeneic immune response observed 7 days after allogeneic transplantation. Furthermore, one of the early signs of skin rejection is epithelial hyperplasia (Lee et al., 2009), and we found that epithelial thickness decreased in allografts on both P2X7-deficient and NLRP3-deficient mice. This was further confirmed by a decrease in the Banff pathological classification for skin allografts (Cendales et al., 2008), where P2X7 or NLRP3 inflammasome deficiency reduced the severity of rejection when assessed 7 days after transplantation. IL-18-deficient mice, but not IL-1 receptor type 1 knockout mice, presented reduced IL-12 production and T cell alloreactivity, thus confirming previous reports where IL-18 has been found to mediate acute rejection in heart and in kidney transplantation (Affleck et al., 2001, Dudler et al., 2007, Striz et al., 2005). Deficiency or antagonism of the P2X7 receptor also resulted in increased skin allograft survival, reduced IFN-γ production, impaired T cell alloreactivity, decreased immune cell infiltration in the transplant, and reduced production of allogeneic antibodies. This suggests that a complete allograft immune rejection response depends on the P2X7 receptor signaling to the NLRP3 inflammasome via the production of IL-18 in sterile conditions, despite the fact that these do not control the early infiltration of macrophages (the present study) or the inflammatory monocytes in the allograft (Oberbarnscheidt et al., 2014).

Potential anti-inflammatory drugs have been developed using drug-like antagonists to target the P2X7 receptor (Arulkumaran et al., 2011, North and Jarvis, 2013). Clinical trials have begun on different selective P2X7 receptor antagonists, with the AstraZeneca AZD9056 compound demonstrating beneficial effects in phase IIa for Crohn’s disease and inflammatory arthritis (Eser et al., 2015, Keystone et al., 2012). In the present study, we also provide evidence that pharmacological targeting of the P2X7 receptor in vivo delays allogeneic graft rejection and increases graft survival, which suggests that this receptor could be a therapeutic target for inducing tolerance without causing severe immunosuppression and supports previous findings that blocking purinergic signaling has a beneficial effect on lung and heart allograft survival (Liu et al., 2014, Vergani et al., 2013, Vergani et al., 2014). P2X7 receptor expression has been associated with disease severity and has been found at higher levels in macrophages associated with coronary disease or rheumatoid arthritis (Al-Shukaili et al., 2008, Stachon et al., 2017, Zanin et al., 2015). Similarly, P2X7 receptor expression increases in the CNS during status epilepticus or optic nerve injury (Jimenez-Pacheco et al., 2013, Nadal-Nicolás et al., 2016). Significantly, P2X7 receptor expression has also been found in mononuclear cells infiltrating allogeneic transplants (Vergani et al., 2013; the present study). Furthermore, we also found an increase in P2X7 receptor expression in inflammatory blood monocytes during human liver transplantation rejection; that is, the rejection severity (measured by the increase in plasma liver enzymes) correlated with the concentration of plasma ATP. Therefore, the increase in P2X7 receptor expression in monocytes during rejection suggests a higher degree of inflammation and, according to our results, a potentially higher risk of graft rejection. In conclusion, our findings show that early ATP released by innate immune cells as a danger signal modulates the activation of the innate and adaptive immune response in allotransplants by activating the NLRP3 inflammasome and IL-18 production.

Experimental Procedures

Human Clinical Samples

All samples were obtained under informed consent. This study was approved by the ethical committee of the Hospital Clínico Universitario Virgen de la Arrixaca (Murcia, Spain, with reference number PI12/02042), in accordance with the ethical guidelines of the 1975 Declaration of Helsinki. Eight liver transplant patients were subjected to progressive immunosuppressive drug withdrawal (Table S1). Four of these patients suffered acute rejection during the immunosuppression withdrawal process (non-tolerant patients), while the other four patients became tolerant and have been living free of immunosuppressive drugs for no less than 1 year (tolerant patients).

Mice Skin Transplantation

Mice were maintained in specific pathogen-free conditions at a room temperature of 22°C ± 2°C and a 12-hr:12-hr light:dark cycle. The mice were fed with a sterile commercial pellet diet and sterile tap water ad libitum. Skin transplantation procedure was approved by Servicio de Sanidad Animal, Dirección General de Ganadería y Pesca, Consejería de Agricultura y Agua, Región de Murcia (reference number C1310050305). Skin transplantation was conducted in accordance with previous protocols (Garrod and Cahalan, 2008). Transplantations were made using male 6- to 8-week-old wild-type C57BL/6 mice or Nlrp3−/−, Casp1/11−/−, P2rx7−/− Il1r1−/−, and Il18−/− mice (all in C57BL/6 background) as receptors (Kuida et al., 1995, Martinon et al., 2006, Solle et al., 2001). Male 6- to 8-week-old BALB/c mice were used as donors in allogeneic transplantations, while C57BL/6 mice were used as donors in isogeneic transplantations. Analgesic and anesthetic levels were carefully monitored every 10 min during the surgical transplant procedure to ensure they were correct. After surgery, animals presenting dehydration were given saline solution. The animals’ well-being was monitored twice per day during the first 7 days after surgery by observing their general aspect and ability to drink, feed, and move normally. No animals had to be sacrificed due to any unexpected significant decrease in animal well-being after surgical procedure.

In Vivo Bioluminescence for Extracellular ATP

106 pmeLuc cells (Morciano et al., 2017, Pellegatti et al., 2005) or 105 HEK293 cells transfected with pLG4.51 were inoculated under the skin graft for 1 hr before recording. Then D-Luciferin was inoculated intraperitoneally (150 mg/kg, 5 min). Imaging was carried out on an IVIS 100 System (Xenogen). Living Image software (PerkinElmer) was used to obtain total photon counts during 15 min of the regions around the skin graft. Apyrase (4 UI) was inoculated under the skin transplant 10 min before imaging.

Skin Transplant Ultrasound

Skin transplants were evaluated daily using the Vevo 3100 Imaging System (Visual Sonics), with a high-frequency probe at 30 MHz. Transversal images of the transplanted area were taken. The infiltrate tissue under the skin graft was measured from the ultrasound image, and the structure of the skin layers was used to determine the survival of the graft as detailed below.

Histopathology

Syngeneic and allogeneic grafts were fixed (4% formalin, 24 hr), processed, paraffin-embedded, and sectioned. H&E staining was used to evaluate the Banff pathological classification for skin allografts (Cendales et al., 2008). The image analysis software Axio Vision (version [v.]4.8.2, Carl Zeiss) was used to measure epidermal thickness by taking the length from the basement membrane to the stratum corneum; at least 10 measurements were taken per single 400× field in 10 fields per sample and from at least 4 different transplants. An indirect ABC-peroxidase was used to detect different immune cell populations in the graft; sections were blocked briefly (3% BSA, 20 min, 37°C) and incubated (overnight, 4°C) with primary antibody (1:500 for anti-CD3, Dako, 1:100 for anti-F4/80, AbD Serotec; 1:300 for CD8, Abcam; and 1:100 for FOXP3, eBioscience). Sections were examined with a Zeiss Axio Scope AX10 microscope with an AxioCamICC3 (Carl Zeiss).

Skin Graft Survival

Ultrasound imaging, histopathological examination, and macroscopic inspection were used to generate an index to evaluate the survival of the skin graft. A graft was considered rejected when it showed the following: (1) hypoechogenic necrotic areas in dermis analyzed by ultrasound and histopathology, (2) abnormal epidermal and dermal constitution detected by ultrasound and histopathology, (3) the presence of eschar formation, and (4) independence of the graft from the surrounding skin. The Banff score was also taken into account.

ELISA

Cytokines were detected in cell supernatants and mice serum samples by ELISA (R&D Systems, Affymetrix, eBioscience, and Cloud Clone), in accordance with the manufacturers’ indications.

Flow Cytometry

Frozen human peripheral blood mononuclear cells (PBMCs) were thawed and dead cells were removed (Miltenyi Biotec). Live cells were incubated with Human BD FC block (BD Biosciences). Cells were stained with anti-human CD14, anti-human CD16 (BD Biosciences), and anti-human P2X7 receptor (clone L4) (Buell et al., 1998). Cells were subjected to flow cytometry analysis in a Cube 8 flow cytometer (Sysmex) The frequency of inflammatory monocytes (CD14+CD16high) and the mean fluorescent intensity of the P2X7 receptor were determined with FCS Express 5 software (DeNovo Software).

qRT-PCR Analysis

qRT-PCR was conducted in accordance with the detailed methods described elsewhere in the literature (Lopez-Castejón et al., 2011).

Statistical Analysis

Statistics were calculated with Prism software (GraphPad). For two-group comparisons, a two-tailed unpaired t test was used. Comparisons of multiple groups were analyzed using one-way or two-way ANOVA with Bonferroni multiple-comparison test. The Gehan-Breslow-Wilcoxon test was used to compare differences in the transplant survival curves. Data are presented as mean ± SEM from the number of independent experiments presented in the figure legends (∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001; ns, no significant difference, p > 0.05).