The control of the maturation and release of the proinflammatory cytokine interleukin (IL)-1β and IL-18 in macrophages by the NLRP3 inflammasome complex is a tightly controlled two-step process. An initial signal primes the production of the immature form of the cytokines and NLRP3. A second signal induces activation of the NLRP3 inflammasome complex with the subsequent activation of caspase-1 and the processing and release into the extracellular space of IL-1β and IL-18 by a pyroptotic process involving the cleavage of gasdermin D [1
]. In human monocytes, an alternative pathway activates the NLRP3 inflammasome without the requirement of a second stimulat [4
]. The second signal necessary to activate NLRP3 inflammasome can be achieved by different stimuli, being the activation of the purinergic P2X receptor 7 (P2X7R) by a concentration of extracellular ATP (eATP) in the mM range among the most common used [6
]. We have recently found that activation of P2X7R before priming signal 1, leads to a mitochondrial damage and a later defect on NLRP3 activation [7
], highlighting the different pathways associated to P2X7R signaling. Extracellular nucleotides such as ATP are indicative of cellular distress, released in small amounts during apoptosis or in faster, larger amounts during sudden cell death like necrosis or pyroptosis [3
]. In this regard, extracellular nucleotides have been implicated in several inflammatory processes either promoting or reducing the inflammatory response depending on the context [10
Depending the type of nucleotide and its concentration, extracellular nucleotides, and their direct degradation products, bind to different purinergic receptors, including adenosine receptors, metabotropic P2Y and ionotropic P2X receptors, providing a wide variety of immune responses depending on the ligand interaction that goes beyond P2X7R activating the NLRP3 inflammasome [11
]. Therefore, it is expected that nucleotides will activate different purinergic receptors expressed in the target cell resulting in different intracellular signaling pathways. This concept led us to hypothesize that initial lower concentration of nucleotides, that do not trigger P2X7R, may regulate NLRP3 inflammasome activity prior its full activation, modulating the overall inflammatory response, beyond the role of P2X7R downmodulating inflammasome response [7
]. To test this hypothesis, we employed murine residential peritoneal macrophages (RPMs), a tissue resident macrophage involved in cell clearance during steady-state conditions as well as in inflammatory responses [12
]. Expression of different purinergic receptors by these cells makes this type of macrophage an excellent model to study the role of nucleotides on NLRP3 inflammasome activation. We found that when RPMs were cultured with nucleotides at μM concentrations that do not trigger P2X7R in the presence of LPS, there was an increase in IL-1β release after NLRP3 activation, and an increase of IL-6 whereas a decrease in TNF-α production in response to the LPS priming. Blockade of purinergic P2Y2
R) reverted IL-1β levels back to amounts obtained only with LPS, whereas decrease production of TNF-α was due to adenosine receptor activity originated from ATP degradation into adenosine. Our data indicates that nucleotides contribute at different levels to a distinct and unique pro-inflammatory signature, which may be important for future anti-inflammatory therapies.
Macrophages are important immune cells to control the initiation of the inflammatory response due to the expression of a wide array of receptors [19
], and therefore are highly sensitive to stimulation with PAMP or DAMP moieties that will induce production of pro-IL-1β cytokine as well as activation of the NLRP3 inflammasome to generate the maturation of this cytokine [20
]. Nucleotides control at different levels the inflammasome-related production of IL-1β and IL-18, however there are reports indicating nucleotides could either induce or inhibit their production [21
]. The idea that nucleotides influence the NLRP3 inflammasome through purinergic receptors other than P2X7 receptor has already been suggested and reviewed [23
], but this hypothesis require better characterization, which prompted us to analyze in detail the effect of different nucleotides on macrophages. In this study, we describe that when murine peritoneal macrophages were primed with endotoxin in the presence of low nucleotide concentrations (in the range of 2–200 μM), such as ATP and/or UTP, there is an induction of IL-1β production through P2Y2
R after NLRP3 activation, whereas caspase-1 activation, IL-18 production, ASC speck formation and pyroptosis remained unchanged. This effect is most probably explained by an increase in Il1b
gene expression induced by P2Y2
R controlling JNK signaling, and not in changes in NLRP3 activity. On the contrary, when extracellular ATP concentration rises to the mM range and P2X7R activates before macrophage priming with endotoxin, we recently found that the activity of NLRP3 inflammasome is reduced [7
Extracellular ATP is able to activate a wide range of purinergic receptors in target cells, and its concentration will dictate the type of receptor activated [25
]. While low concentrations of ATP activate P2Y receptors and P2X1–6 receptors, higher concentrations are needed to activate P2X7R [25
]. P2X7R is linked to different signaling pathways, and its activation could contribute not only to the activation of the NLRP3 inflammasome, but also affects the cellular energy metabolism, host-pathogen interactions and cell death [26
Interestingly, other inflammasome-independent proinflammatory cytokines, like IL-6 and TNF-α, were increased and decreased respectively, indicating a specific inflammatory response of macrophages when confronted to PAMPs and nucleotides. Decrease of TNF-α has already been described when adenosine receptors are activated together with LPS [10
], and our study also confirmed adenosine receptors as responsible for TNF-α decrease, probably due to ATP degradation to adenosine. The fact that nucleotides imprint a specific inflammatory signature in macrophages has important consequences for anti-inflammatory treatment therapies in diseases such as inflammatory bowel disease, autoimmune arthritis, cardiovascular diseases, cancer or even obesity [28
]. IL-6 is found increased after nucleotide treatment, suggesting that P2Y2
R could favor the release of cytokines, similarly to the effect of P2Y2
R found to be able to release MCP-1 without altering the mRNA levels [14
Our study found P2X4 receptor as highly expressed in mouse peritoneal macrophages, as has been previously reported [32
]. However, the relatively high expression of p2yr2
gene as well as response to low concentrations of ATP or UTP directed us to target P2Y2
R as the sensor responsible for the observed increase in IL-1β. Nonetheless, we cannot rule out a combined effect of different purinergic receptors affecting the increase of IL-1β, whose effect is reduced if one of the receptor’s signaling, as P2Y2
R, is absent or blocked. In fact, P2X4 receptor has been implicated in IL-1β release [33
] and could also modulate the increase of IL-1β when ATP was applied to the macrophages. We found that the inhibition or genetic deficiency of P2Y2
R restored NLRP3-dependent IL-1β release to levels obtained only with LPS priming. Given that P2Y2
R signaling affects pro-IL-1β synthesis, the increase in IL-1β production should not be restricted to the activation of NLRP3, but also could affect other inflammasome activation, such as NLRC4. However, the differential presence of P2Y2
R in different type of macrophages will shape the response to ATP or UTP, since this effect is not present in THP1 or BMDM. This could be due to either a lack of P2Y2
R receptor or membrane expression or P2Y2
R might not be similarly coupled to JNK signaling and differently coupled to other pathways. The role of P2Y2
R is unclear in the inflammatory processes, with paradoxical reports. P2Y2
R has been involved in cell clearance processes and thus helping maintaining homeostasis [8
]. However, P2Y2
R has also been shown to be involved in pro-inflammatory responses [35
]. Recently, it has been described that P2Y2
R is required to induce IL-1β production in irradiated tumor cells by a pannexin-1-dependent mechanism [37
], although the downstream P2Y2
R signaling is unknown, it would suggest a physiological context in which macrophages would show a pro-inflammatory response as indicated in this manuscript. P2Y2
R, as many G protein-coupled receptors, signals through PI3K, PLC activation and a subsequent intracellular calcium release [14
], although the exact mechanism remains unclear [38
]. However, we have not observed a requirement for PI3K, PLC or intracellular calcium signaling when P2Y2
R increased IL-1β production, suggesting an alternative pathway of the classical P2Y2
R activation, as it has been described for endothelial cells [40
]. Similarly, PLC-independent mechanisms by which P2Y2
R can modify cytokine production in macrophages have also been described [14
], although the exact mechanism remains unclear.
The observation that Il1b
gene expression was increased by P2Y2
R activation prompt us to test other pathways that were also downstream purinergic receptors and induced Il1b
gene transcription, and MAPK have been recently found downstream purinergic receptors in human monocytes [41
]. We found that inhibition of JNK MAPK resulted in a decrease in nucleotide-induced IL-1β, whereas inhibition of the other two classical MAPK, ERK1/2 and p38, resulted in a general IL-1β decrease, but did not affect P2Y2
R-increased IL-1β release. The importance of JNK activity in IL-1β production by macrophages has already been described for the activation of the inflammasome by calcium crystals [43
] or palmitate [44
]. Increased NLRP3 activation by ATP or UTP has also been described; however, P2Y2
R was not responsible for this activation [45
]. Furthermore, JNK has been implicated in the direct activation of NLRP3 inflammasome by phosphorylating its PYD domain and favoring ASC engagement [46
]. Our results confirm that P2Y2
R was not affecting the activation of NLRP3, but enhanced Il1b
transcription by JNK activation, and this model is not incompatible with a direct NLRP3 phosphorylation.
We also describe how nucleotides influence macrophage response depending on cell density. A common approach used in experiments to assess inflammasome activation is to set cell cultures at full confluence. However, macrophages in vivo are present more dispersed in steady state conditions or even during pathological settings, with some exceptions as parasite infections, foreign object presence, granulomas and some tumor locations, where they aggregate at high concentration. It is therefore of interest to analyze the function of macrophages at low densities. We have found that IL-1β production can be enhanced by nucleotides only when macrophages are cultured at low concentration. This effect correlates with an increased JNK activation in cells at high concentration. Signaling through JNK (originally named stress activated protein kinase) in macrophages is linked to inflammation [47
]. High cell density cultures provide neighboring cells with important interactions that resulted in an important increase in MAPK activity (and probably other pathways) and possible more stable signals sustained in time than the activation of P2Y2
R. As a consequence, much higher IL-1β production is obtained from high density cultures, even when high density is reached supplementing NLRP3-deficient macrophages to low density wild type macrophages.
In summary, future studies to further examine the role of extracellular nucleotides and P2Y2R signaling during in vivo inflammatory conditions, as well as their potential as novel receptors to treat inflammation are warranted.
4. Materials and Methods
ATP, UTP, adenosine 5′-o-(3-thiotriphosphate) tetralithium salt (ATP-γS), and nigericin were purchased from Sigma-Aldrich (St. Louis, MO, USA). AR-C118925xx, SP600125, SB202190, U0126, U73122, SCH58261, MRS1754, BAPTA-AM, thapsigargin, wortmannin were purchased from Tocris Bioscience (Bio-techne, Bristol, UK).
(wild-type) mice were purchased from Harlan Laboratories (Indianapolis, IN, USA) and bred in the local animal facility. NLRP3-, Caspase-1/11- and P2X7R-deficient (Nlrp3−/−
] mice in C57BL/6
background were bred in our facilities. For all experiments, mice aged 8–24 weeks were used, in accordance with the University Hospital Virgen Arrixaca animal experimentation guidelines, and the Spanish national (RD 1201/2005 and Law 32/2007) and European Union (86/609/EEC and 2010/63/EU) legislation. According to the cited legislation, local ethics committee review or approval is not needed, because the mice were killed by CO2
inhalation and used to obtain peritoneal lavage or tissues. No procedure was undertaken to live animals that compromised animal welfare.
4.3. Isolation and Culture of Macrophages
Resident peritoneal macrophages (RPMs) were isolated from resting C57/BL6 mice peritoneal cavity previously euthanized with CO2
by lavage using cold PBS with 2 mM EDTA. Macrophages were further enriched by magnetic depletion of CD19+
cells using magnetic microbeads (Miltenyi Biotech, Bergisch Gladbach, Germany). To confirm macrophage enrichment, initial tests of samples before and after magnetic bead purification were analyzed by flow cytometry for expression of F4/80 (antiF4/80-alexa488, clone BM8, BioLegend, San Diego, CA, USA), MHC-II (anti-MHC-PE conjugated, clone M5/114.15.2, eBiosciences, San Diego, CA, USA) or CD19 (PE-conjugated, clone eBio1D3, eBiosciences) and analyzed in a FACSCanto cytometer (BD Biosciences, San Diego, CA, USA). Cells were plated at 0.3 × 106
cell/mL (low density) or 1.2 × 106
cell/mL (high density) for 1 h in RPMI-1640 media (Life Technologies, Carlsbad, CA, USA) with 10 mM HEPES and 2 mM L-glutamine (BioWhittaker—Lonza, Basel, Switzerland) (supplemented RPMI) and complemented with 5% fetal bovine serum (FBS, Life Technologies). Plate wells were then rinsed with pre-warmed PBS to remove non-adherent cells and further enrich the macrophage culture. Cells were then cultured for a minimum of 2 h in supplemented RPMI with 0.5% endotoxin free, sterile filtered bovine serum albumin (Sigma). Bone marrow derived macrophages (BMDM) were differentiated for 7 days in the presence of L-cell media as already described [50
]. Differentiation of THP-1 cells was performed in RPMI media supplemented with 10% FBS and 0.2 μM PMA for 4 h, then media was replaced with fresh media with FBS without PMA and cells were incubated overnight. Cells were then rinsed and media without FBS was added for cell stimulation.
Cells were stimulated for 3 h with either 200 ng/mL of ultrapure LPS from E. coli 0111:B (InvivoGen, San Diego, CA, USA), 2 μg/mL of Pam3CSK4 (InvivoGen) or 20 μg/mL of Poly I:C (InvivoGen) in the presence or absence of 20 μM of ATP and/or 20 μM of UTP, unless otherwise indicated. Supernatants from this initial priming step were recovered when needed for cytokine detection, and plates were rinsed with PBS and then with physiological buffer (147 NaCl, 10 HEPES, 13 D-glucose, 2 KCl, 2 CaCl2, and 1 MgCl2; pH 7.4, all in mM concentration). Finally, cells were stimulated with physiological buffer containing or not 3 mM of ATP or 5 μM nigericin, for 24 min, and then supernatants were recovered, cleared and stored at −80 °C.
4.4. LDH Determination
Pyroptosis was analyzed by measurement of released lactate dehydrogenase (LDH) in the supernatants using the Cytotoxicity Detection kit (Roche, Barcelona, Spain) following the manufacturer’s instructions, and expressed as percentage of total cell LDH content, using samples from cells lysed in 1% Triton X-100 buffer.
Supernatants from cultured macrophages in duplicate wells were cleared at 500× g to remove any remaining cell. IL-1β was analyzed with Affimetrix’ Ready-Set-Go (BioLegend) ELISA kit. TNF-α, IL-6 and IL-18 were analyzed by Quantikine ELISA (R&D, Biotechne, Minneapolis, MN, USA).
RPMs were seeded at the desired concentration (0.4 × 106 or 1.2 × 106 cell/mL) onto coverslips with RPMI media with 5% FBS. After being activated with 5 μM nigericin, 3 mM ATP, cells were fixed in 2% paraformaldehyde. Cells were blocked with autologous serum and stained with primary anti-ASC (HASC-71, BioLegend) and secondary donkey anti-mouse AlexaFluor488 (Molecular Probes, Thermo Fisher Scientific, Waltham, MA, USA) and mounted on slides with DAPI-containing mounting medium (Prolong diamond antifade, Life Technologies). Images were acquired with an Eclipse Ti microscope (Nikon, Tokyo, Japan) equipped with a 10× (numerical aperture, 0.30) or 20× S Plan Fluor objective (numerical aperture, 0.45) and a digital Sight DS-QiMc camera (Nikon) and 387 nm/447 nm and 482 nm/536 nm filter sets (Semrock, Lake Forest, IL, USA).
4.7. Quantitative Reverse Transcriptase-PCR Analysis
mRNA was obtained using the RNeasy Mini kit (Qiagen, Venlo, The Netherlands) as per manufacturer instructions. Quantitative PCR was performed using SYBR Premix ExTaq (Takara, Göteborg, Sweden). Specific primers were purchased from Qiagen (QuantiTech Primer Assays). For each primer set, the efficiency was >95%, and a single product was obtained on melt curve analysis. The presented relative gene expression levels were calculated using the 2ΔΔCt method normalizing to the endogenous Hprt1 expression levels, as a house keeping control, for each treatment, and the fold increase in expression was relative to the smallest expression level or to the control basal levels.
4.8. Western Blot Analysis
Cells were plated at 0.3 × 106 cell/mL (low density) or 1.2 × 106 cell/mL (high density) were rinsed with cold PBS and lysed in 1% NP40 buffer supplemented with protease inhibitor cocktail (Sigma-Aldrich) and phosphatase inhibitor (PhosSTOP, Roche). Cell lysates and supernatants were resolved in 12% acrylamide SDS-PAGE gels and blotted into a PVDF membrane for mIL-1β (H-153, Santa Cruz, Dallas, TX, USA), NLRP3 (Cryo-2 AG-20B-0014, Adipogen, Liestal Switzerland) or anti MAPK antibodies anti-Phospho JNK (Thr183,Tyr 185, Cat. 9251S), anti-JNK (Cat. 9252S), anti-Phospho-p44/42 (phospho-Erk1/2) (Thr202, Tyr 204, Cat. 4377), anti-p44/42 (Erk1/2), anti-phospho p38 (Thr 180/Tyr 185, Cat. 9211) or anti-p38 (Cat. 9212) (all from Cell Signaling Technology, Danvers, MA, USA). Primary antibody incubation was performed overnight in 3% bovine serum albumin (Sigma-Aldrich) or 5% w/v Difco skim milk (BD Biosciences). Primary antibodies were revealed using the corresponding secondary anti-mouse, or anti-rabbit IgG-peroxidase horseradish linked (GE-Healthcare, Munich, Germany). Analysis of protein bands was performed using Image Lab software (Bio-Rad Laboratories, Hercules, CA, USA), and values were normalized to β-actin.
4.9. Statistical Analysis
Data is shown with +SEM or +SD as indicated. For some experiments (mainly ELISAs) data was normalized dividing tests values by that obtained by control (i.e., LPS only, or no inhibitor control), thus, control test was given value “1” and the tests results are proportional to that control. In other experiments, due to large differences in the potency of the response of the cells in the different repeated experiments, we show a representative experiment of the replicates. Statistical analysis was performed using Prism software (GraphPad Inc., La Jolla, CA, USA) by testing two-way ANOVA with Sidak’s multiple comparison test for multiple comparisons or by multiple t-test between 2 groups.