A₁ and A_2A adenosine receptors play a protective role to reduce prevalence of autoimmunity following tissue damage

Summary

Adenosine is a potent modulator that has a tremendous effect on the immune system. Adenosine affects T cell activity, and is necessary in maintaining the T helper/regulatory T cell (T_reg) ratio. Adenosine signalling is also involved in activating neutrophils and the formation of neutrophil extracellular traps (NETs), which has been linked to autoimmune disorders. Therefore, adenosine, through its receptors, is extremely important in maintaining homeostasis and involved in the development of autoimmune diseases. In this study, we aim to evaluate the role of adenosine A₁ and A_2A receptors in involvement of autoimmune diseases. We studied adenosine regulation by NETosis in vitro, and used two murine models of autoimmune diseases: type I diabetes mellitus (T1DM) induced by low‐dose streptozotocin and pristane‐induced systemic lupus erythematosus (SLE). We have found that A₁R enhances and A_2AR suppresses NETosis. In addition, in both models, A₁R‐knock‐out (KO) mice were predisposed to the development of autoimmunity. In the SLE model in wild‐type (WT) mice we observed a decline of A₁R mRNA levels 6 h after pristane injection that was parallel to lymphocyte reduction. Following pristane, 43% of A₁R‐KO mice suffered from lupus‐like disease while WT mice remained without any sign of disease at 36 weeks. In WT mice, at 10 days A_2AR mRNA levels were significantly higher compared to A1R‐KO mice. Similar to SLE, in the T1DM model the presence of A₁R and A_2AR was protective. Our data suggest that, in autoimmune diseases, the acute elimination of lymphocytes and reduction of DNA release due to NETosis depends upon A₁R desensitization and long‐term suppression of A_2AR.

The role of adenosine A₁ and A_2A receptors in the development of autoimmune diseases. In both diabetes and lupus models, A₁R‐KO mice were predisposed to the development of autoimmunity. We also found that A₁R enhances and A_2AR suppresses NETosis. Our data suggest that in autoimmune diseases, the acute elimination of lymphocytes and reduction of DNA release, due to NETosis, depends on A₁R desensitization and long‐term suppression of A_2AR.

INTRODUCTION

Autoimmune diseases are chronic conditions characterized by the loss of immunological tolerance to self‐antigens, leading to self‐attack of the immune system on the organs. Genetic, environmental, hormonal and immunological factors are considered important in the etiology of autoimmune diseases such as systemic lupus erythematosus (SLE) (1), rheumatoid arthritis (RA) (2), type I diabetes mellitus (T1DM), etc. (3, 4, 5). In the course of these diseases we are witnessing the involvement of various immune cell types. For instance, T cells are capable not only serving as inducers of autoimmune disease but also to inhibit such disease. Naive CD4⁺ cells can differentiate into T helper cells that secrete cytokines, but in the presence of tumor necrosis factor (TNF)‐β will differentiate into T regulatory cells (T_regs) (reviewed in (6)), a special population of T cells that modulate the immune system and maintain tolerance to self‐antigens. Therefore, T cells are extremely important in maintaining homeostasis and are involved in the development of autoimmune diseases (7). Neutrophils are also significant regulators of immune responses, which are found to be key players in various autoimmune diseases (8). Defective development, growth and death of neutrophils may lead to the initiation of autoimmune diseases (reviewed in (8)). Another potent immune‐modulator that has a tremendous effect on the immune cells during autoimmunity is adenosine (9, 10, 11). Adenosine can affect T cell development, proliferation and activity (12, 13, 14). Adenosine acts on four receptors, all of which are members of the G protein‐coupled receptor family: A₁ receptor (A₁R), A_2A receptor (A_2AR), A_2B receptor (A_2BR) and A₃ receptor (A₃R). The A₁R and A₃R activate Gi protein, which inhibits adenylyl cyclase activity and decreases cyclic adenosine monophosphate (cAMP) levels and promotes a proinflammatory response. A_2AR interacts with Gs and the A_2BR interacts with Gs/Gq to induce adenylyl cyclase activity and elevates cAMP levels, thus promoting anti‐inflammatory responses (15). Adenosine can be generated from adenosine triphosphate (ATP) degradation by ecto‐nucleoside triphosphate diphosphohydrolase 1 (CD39), which catalyzes the sequential transition of ATP to adenosine diphosphate ADP and adenosine monophosphate (AMP) and by the ecto‐5′‐nucleotidase (CD73), which dephosphorylates AMP. Both CD39 and CD73 are highly expressed on T_regs and regulatory B cells (B_regs), and are therefore important for shifting the balance between inflammation and immunosuppressive environment (16, 17). The adenosine signaling pathway is also necessary for T_regs to exert their suppressive function. A_2AR has a critical role in maintaining T helper/T_regs ratio, as well as the overall T/B lymphocyte ratio within the germinal centers in secondary lymphoid organs (18).

Adenosine also regulates the activation of neutrophils (19). A₁R promotes neutrophil chemotaxis, whereas A_2AR and A_2BR inhibit neutrophil activation (20). Adenosine signaling is also involved in the formation of the so‐called neutrophil extracellular traps (NETs), which are chromatin filaments coated with proinflammatory and effector molecules released by neutrophils into the extracellular space in response to inflammatory triggers. NETs formation has the main role of containing pathogen spreading, but it has been also linked to autoimmune disorders. Activation of A_2AR reduces NETs formation, thus contributing to down‐regulating NET proinflammatory activity and counteracting the development of autoimmunity (21).

The critical involvement of adenosine in autoimmune diseases is supported by many studies (reviewed in (22)). For example, in a murine model of lupus, it has been shown that A_2AR agonist results in significant improvements in renal function (23). In children with T1DM it has been observed that low CD39 levels correlate with disease activity, suggesting a potential compromise in T_reg function (24). Abnormal regulatory immune modulators are critical to the pathogenesis of autoimmune diseases, therefore adenosine and the dynamics of its receptors are important to understand during such diseases. We have previously shown in a model of systemic inflammatory response syndrome (SIRS) that a surge of adenosine desensitizes Gi‐coupled adenosine A₁R and up‐regulates Gs‐coupled A_2AR, which is an important part of the pathology in SIRS and probably needs to be taken into consideration while treating (25). We hypothesize that following traumatic events the suppression of the immune system associated with elevation of A_2AR prevents autoimmunity. Therefore, in the current study, we aimed to assess the roles that A₁R and A_2AR have in the course of autoimmune diseases in two important murine autoimmunity models (lupus and T1DM).

MATERIALS AND METHODS

Mice

Experiments were conducted after obtaining permission from the Israel Committee for Animal Experiments [interleukin (IL)‐32‐06‐2013, IL‐25‐5‐2016, IL‐39‐8‐2017]. BALB/c and C57BL/6 mice were purchased from Harlan (Jerusalem, Israel), A₁R knock‐out (KO) mice (A₁R^−/− on a C57BL/6 background) and A_2AR knock‐out mice (A_2AR^−/− on a BALB/c background) were purchased from the Jackson Laboratory (Bar Harbor, ME, USA). Mice were housed under specific pathogen‐free conditions and maintained in the vivarium of Ben‐Gurion University. All experiments were approved by the Ben‐Gurion University Committee for Ethical Care and Use of Animals in Experiments.

Agonists, antagonists and inhibitors

A₁R agonist 2‐chloro‐N⁶‐cyclopentyladenosine (CCPA) and A_2AR agonist 2‐p‐(carboxyethyl) phenethylamino‐5’‐N‐ethylcarboxamideadenosine hydrochloride (CGS21680), were purchased from Sigma‐Aldrich (Rehovot, Israel).

Regulation of NETosis by adenosine

Differentiated human leukemia (HL)‐60 cells

HL‐60 cells (CCL240; American Type Culture Collection) were grown in RPMI‐1640 and supplemented with 10% heat‐inactivated fetal calf serum (FCS), 2 mmol/l L‐glutamine, 100 U/ml penicillin and 100 μg/ml streptomycin (Biological Industries, Bet Haemek, Israel). HL‐60 cells were differentiated into neutrophils by culturing the cells in medium containing 5 µM retinoic acid (RA; Sigma‐Aldrich) for 72 h. Differentiation was confirmed by detection of surface CD11b, which is an early marker of neutrophil differentiation in HL‐60 (26). Differentiation was confirmed when CD11b expression levels were at least 90%. Untreated HL‐60 cells stained with the isotype control were used as background for undifferentiated cells.

NETs assay

RA‐differentiated HL‐60 neutrophils or bone marrow (BM)‐isolated cells were seeded (2 × 10⁵ per well) in 96‐well plates. Cells were preincubated with or without adenosine agonists (A₁R‐specific agonist CCPA, 1 nM or A_2AR agonist GCS, 30 nM) for 30 min. They were then treated with 200 nM phorbol 12‐myristate 13‐acetate (PMA; Sigma‐Aldrich) and 0.03% H₂O₂ for 3 h (27, 28). For DNA detection, Sytox green dye (Molecular Probes, Invitrogen AG; Basel, Switzerland) was used.

Type 1 diabetes model (TID)

We employed the model of low‐dose streptozotocin (STZ; Sigma‐Aldrich) to induced T1DM in all our experiments. In this model, diabetes develops only when STZ induces both β cell toxicity and T cell‐dependent immune reactions (29). We employed a regimen involving multiple administrations of low‐dose STZ in mice (30). Diabetes was induced in 8‐week‐old C57BL/6 mice of both sexes by intraperitoneal (i.p.) injection of STZ (50 mg/kg in citrate buffer) on 5 consecutive days. Blood glucose levels were measured using a glucometer (Accu‐Chek Aviva, Roche Diagnostics, Indianapolis, IN, USA). Regularly, in all STZ‐injected mice throughout the experiment, animals with glucose levels >200 mg/dl for 2 consecutive days were considered to be diabetic (31).

Lupus model

Pristane, a natural saturated terpenoid alkane obtained primarily from shark liver oil, was shown to induce a lupus‐like disease in mice (32). Injection of pristane into the peritoneal cavity results in chronic peritonitis associated with high tissue levels of IL‐6 (33), which leads to a slow process to lupus‐like disease (34).

Mice were injected i.p. with 0.5 ml of pristane (Sigma Aldrich) and followed weekly for external signs of lupus such as alopecia, chronic wounds or death. Spleen, blood and peritoneal lavage were collected at euthanasia (6, 24, 48, 6 and 10 days or 8 months after injection) (35).

Anti‐double‐strand DNA (dsDNA)

Disease activity was considered according to anti‐dsDNA antibodies (36). Serum was separated by centrifugation at 4°C at 1000 g for 10 min, and serum anti‐dsDNA levels were analyzed using a murine ds‐DNA standard enzyme‐linked immunosorbent assay (ELISA) kit (Alpha Diagnostics, Inc., San Antonio, TX, USA).

Differential blood cell counts

Blood samples of 200 µl in heparin‐coated tubes were counted with an ADIVA 2120 blood count device (Siemens, Munich, Germany).

Cell‐free DNA assay

Peritoneal lavage was performed with 5 ml of phosphate‐buffered saline (PBS) at the experiment end‐point. cfDNA was quantified, as previously described, using our rapid SYBR^® Gold fluorometric assay (37).

mRNA analysis by quantitative polymerase chain reaction(qPCR)

Splenocytes production

For mRNA levels, spleens were harvested, and cells were collected and treated with red blood cell (RBC) lysis solution (5 Prime, Inc., San Francisco, CA, USA). Cells were incubated in a Petri dish at 37ºC with medium for 1 h. They were then washed, adhered cells were collected and RNA was extracted using a PerfectPure RNA Tissue Kit (5 Prime, Inc.).

RNA was extracted using a Perfect Pure RNA Tissue Kit (5 Prime, Inc.). cDNA was prepared using a high‐capacity cDNA reverse transcription kit (Applied Biosystems, Foster City, CA, USA).

qPCR assays were performed with a Fast SYBR Green Master Mix (Applied Biosystems) on a StepOne Plus real‐time PCR machine (Applied Biosystems) with the following mouse‐specific primers: RPL‐12 sense 5′‐ATG ACA TTG CCA AGG CTA CC‐3′, anti‐sense 5′‐CAA GAC CGG TGT CTC ATC TGC‐3′; A₁R sense 5′‐TAC ATC TCG GCC TTC CAG GTC G‐3′, anti‐sense 5′‐AAG GAT GGC CAG TGG GAT GAC CAG‐3; A_2A sense 5′‐ CGC AGG TCT TTG TGG AGT TC‐3, anti‐sense 5'‐TGG CTT GGT GAC GGG TATG‐3'.

Statistical analysis

All comparisons between groups were carried out using a Mann–Whitney non‐parametric t‐test or one‐way analysis of variance (anova) followed by a Tukey post‐test using Prism version 6 software (GraphPad, San Diego, CA, USA). P‐values below 0.05 were considered significant. Data are presented as mean ± standard deviation (s.d.) unless mentioned otherwise.

RESULTS

DNA released from neutrophils that undergo NETosis is a major source of cfDNA and underpins the progression of autoimmune diseases (36). To explore the regulation of NETosis by adenosine receptors, we used differentiated HL‐60 cells stimulated for NETosis by PMA + H₂O₂. Stimulation of neutrophil‐like HL‐60 cells with CCPA (1 nM), a specific A₁R agonist before induction of NETosis, increased NETs production while pretreatment with the A_2AR agonist CGS21680 (30 nM) diminished NETs production (Figure 1).

FIGURE 1.

Adenosine receptors regulate neutrophil extracellular traps (NET) production. (a) Human leukemia (HL)‐60 cells differentiated by retinoic acid (RA) to neutrophil‐like cells (2 × 10⁵ cells/well) in triplicate were pre‐exposed to A₁ adenosine receptor agonist [2‐chloro‐N⁶‐cyclopentyladenosine (CCPA), 1 nM] or A_2A adenosine receptor agonist [phenethylamino‐5′‐N‐ethylcarboxamideadenosine hydrochloride (GCS), 30 nM], and cells were then stimulated with phorbol myristate acetate (PMA) (200 nM) and H₂0₂ (0.03%) in the presence of DNA fluorescent dye (5 µM SYTOX green) for 3 h. Relative NETs production was measured by fluorescence in 96‐well plates and normalized to NETs production by cells treated with only H₂O₂ + PMA.*P < 0.05, **P < 0.01. Values are mean ± standard error (s.e.). At least two independent experiments were performed

A₁ and A_2A adenosine receptors have a different and opposite role in inflammation progress (38); nevertheless, absence of both receptors were found to accelerate the induction of T1DM. At 15 days after the first STZ injection, all A₁R‐KO mice were diabetic (glucose > 200 mg/dl), while the wild‐type (WT) (C57 black background) group at this time‐point remained free of disease (Figure 2a). Similarly, A_2AR‐KO mice (BALB/c background) were also more susceptible than WT mice to the early development of T1DM. On day 20 all A_2AR‐KO mice had elevated blood glucose levels, while only 40% of WT mice propagated T1DM (Figure 2b). In agreement with our results, murine models of A_2AR‐KO develop stronger autoimmune disease in a shorter time (23, 36, 39). However, to investigate whether A₁R has a role in autoimmunity or specifically in the development of T1DM, we performed another model. Similarly to our T1DM observation, we observed the appearance of pristane‐induced lupus‐like disease only in A₁R‐KO mice compared to WT mice. A₁R‐KO mice began to exhibit the classic pathological signs of lupus: alopecia, chronic skin wounds and death starting long before the WT mice. At 36 weeks following pristane injection, 43% of A₁R‐KO mice suffered from lupus‐like disease; five died, three suffered from alopecia and one suffered from chronic wounds, while WT mice had no physical signs of disease (Figure 3a), although they developed anti‐dsDNA (Figure 3b), which is a hallmark of lupus. Spleens were removed and measured at the experiment end‐point and were significantly larger following pristane injection in WT mice (Figure 3c).

FIGURE 2.

Adenosine receptors and susceptibility to induced autoimmune type I diabetes. Mice were injected with low‐dose streptozotocin (STZ, 50 mg/kg) for 5 consecutive days. Mice were considered diabetic when glucose remained above 200 mg/dl. The experiment was ended when, in one of the groups, all animals were sick. (a) C57BL/6 wild‐type (WT) and A₁ receptor‐knock‐out (A₁R‐KO) mice, (b) Balb/C WT and A_2AR‐KO mice. Mantel–Cox test signs of disease graphs (n = 5–7)

FIGURE 3.

Susceptibility of A₁ receptor‐knock‐out (A₁R‐KO) mice to pristane‐induced lupus (PIL). C57BL/6 wild‐type (WT) and A₁R‐KO mice were injected with pristane and monitored for alopecia, chronic wounds or death for 36 weeks. (a) Rate of disease appearance. *P < 0.05 (n = 8–14). Mantel–Cox test signs of disease graphs. After 36 weeks, mice were euthanized and analyzed for (b) anti‐dsDNA levels in serum of surviving mice by enzyme‐linked immunosorbent assay (ELISA) and (c) spleen size of surviving mice was measured. *P < 0.05, **P < 0.01 (n = 3–8). Values are mean ± standard error (s.e.)

These results, although unexpected, follow our results in the T1DM model. Thus, we evaluated the short‐term period after pristane injection using complete blood counts. The basal white blood cell (WBC) counts were lower in A₁R‐KO mice compared to WT mice (Figure 4a). In blood counts of WT mice following pristane injection, we observed a sharp reduction in WBC count, where the main leukocyte population to be affected was the lymphocyte count (Figure 4b). At first, lymphocyte counts were reduced and partial recovery was observed 48 h after injection. Neutrophil counts reduced significantly in A₁R‐KO mice following pristane injection compare to T = 0 (Figure 4c).

FIGURE 4.

Lymphopenia following pristane injection. Following pristane injection to C57BL/6, wild‐type (WT) mice and A₁ receptor‐knock‐out (A₁R‐KO) mice (n = 3–6). Blood counts were performed at the indicated time‐points. (a) White blood cells (WBC), (b) lymphocytes, (c) neutrophils and (d) monocytes. *P < 0.05, **P < 0.01, compared to control (WT at t = 0). Values are mean ± standard error (s.e.)

We followed the A₁R and A_2AR mRNA levels (normalized to RPL‐12) at several time‐points for 10 days (Figure 5). In WT mice, parallel to the decline of WBC, A₁R mRNA levels dropped 6 h after pristane injection (Figure 5a) and returned to basal levels 24 h after treatment. In these mice A_2AR was induced following injection, reaching significance at 10 days (240 h, Figure 5b). In contrast, A₁R‐deficient mice failed to up‐regulate the immunosuppressive receptor A_2AR and stayed low for 10 days (Figure 5b).

FIGURE 5.

A₁ receptor (A₁R), A₁R and A_2AR mRNA levels in pristane‐induced lupus (PIL). Pristane was injected into C57BL/6 wild‐type (WT) and A₁R‐knock‐out (KO) mice. To examine the dynamic expression of the two high‐affinity adenosine receptors, A₁R and A_2AR, the spleen was removed at the indicated time‐points (6, 24 and 48 h and 10 days). A₁R and A_2AR mRNA levels in adherent splenocytes were analyzed by real‐time polymerase chain reaction (PCR) and normalized by housekeeping ribosomal protein L12 (RPL‐12) levels. Results are median + interquartile range (n = 3–6). *P < 0.05, between expression levels of each receptor to expression at time 0. ^P < 0.5 compared to WT at the same time‐point. At least two independent experiments were performed

As mentioned above, cell‐free DNA (cfDNA) is a product of neutrophils undergoing NETosis, and is considered to be a marker of the progression of autoimmune diseases (36). In both models, baseline cfDNA levels were elevated in A1R‐KO mice compared to WT mice. Also, in both models, cfDNA levels were elevated in the autoimmune group compared to untreated mice (Figure 6).

FIGURE 6.

cfDNA in autoimmune diseases. Peritoneal lavage was collected to examine the levels of cfDNA in (a) the pristane‐induced lupus (PIL) model and (b) the type I diabetes mellitus (T1DM) model at day 10. After the first injection, analyzed for cfDNA levels by a direct rapid fluorometric assay with the fluorochrome SYBR Gold (lower panel). Results are median + interquartile range, (n = 3–6). P < 0.05, **P < 0.01, ***P < 0.001, between expression levels of each receptor to expression at time 0. ^^P < 0.01, ^^^P < 0.001 compared to WT at the same time‐point (n = 5–7)

DISCUSSION

Multiple reports suggest that the onset of autoimmune disorders is related at least in part to a partial or complete loss of function in the purinergic pathways and local defective production of adenosine (reviewed in (22)).

In the present study, we suggest that adenosine regulates the release of DNA by NETosis and that the same A₁R/A_2AR‐dependent immunosuppressive mechanism reduces cfDNA levels.

Although differing from each other, autoimmune diseases present several shared common phenotypes: high levels of cytokines, the presence of infiltrating immune cells and the presence of non‐specific autoantibodies; i.e. anti‐nuclear antibodies and dsDNA. The latter is a known hallmark of lupus and other autoimmune diseases, but it is not only a disease marker: it also promotes autoimmunity. The presence of dsDNA in the cytoplasm has been described as a potent danger signal that activates the stimulator of interferon genes (STING), a regulator of the immune response (40). Activating STING leads to a signaling cascade that eventually alters proinflammatory molecule production. A defect or unnecessary alert in this mechanism has been described as underpinning the autoinflammatory process (41). One source of dsDNA is related to the production of NETs: activated neutrophils that extrude their DNA and bactericidal molecules, creating NETs in a unique type of cell death called NETosis (15). Neutrophils from patients with various autoimmune diseases are more likely to undergo NETosis than those of healthy donors (42). Moreover, the presence of autoantibodies promotes the release of NETs.

We studied the regulation of NETosis by agonists of adenosine receptors. Similar to the effect on other neutrophil functions, such as adherence to endothelium, chemotaxis, activation and trafficking (8, 9, 10, 11, 12), the A₁R agonist also enhanced NETs production. A_2AR was found to be a negative regulator of NETosis when stimulation of cells with a specific A_2AR agonist decreased NETs production of untreated cells. In support of our data, Liu et al. showed that activation of A_2AR inhibits neutrophil cell death, and suggest that this finding is part of the anti‐inflammatory role of A_2AR in modulating neutrophil survival during SIRS (43). Moreover, a recent study by Ali et al. has shown that the A_2AR agonist attenuates NETosis in anti‐phospholipid syndrome (44).

As A_2AR is known for its anti‐inflammatory activity, A_2AR‐KO mice are expected to develop autoimmune disease faster and worsen compared to WT mice. In accordance with our expectations, mice were rapidly ill, as shown in our T1DM model and as shown in the work by Zhang et al. in mice with lupus nephritis (23). In recent studies, Patinha et al. used the model of STZ in rats and also showed that agonist for A_2AR has a protective role in kidneys in hypertensive diabetic nephropathy (45, 46). T1DM was shown to modify the expression of adenosine receptors in the brain and alteration in the balance of A₁R and A_2AR also effects locomotor activity (47). A_2AR seems to up‐regulate in coronary flow during diabetes development (48), and the activity of A_2AR is needed for the therapeutic activity of Mg2⁺ (49). While the role of A_2AR is not completely understood it is undoubtedly important, and requires further study.

Surprisingly, our data from mice with lupus‐like disease and T1DM show that autoimmune diseases are also exacerbated without the presence of A₁R. Similar to our study, Tsutsui et al. showed in experimental allergic encephalomyelitis that, compared to WT mice, A₁R‐KO mice developed a severe progressive–relapsing form of the disease (50). In our lupus model, we observed that differences following pristane injection in anti‐dsDNA and spleen size did not reach significance between WT and A₁R‐KO mice, as the mice with the most severe illness died before the experiment end‐point and therefore are not included in these results. Following pristane injection, we observed an acute reduction of A₁R and leukocyte counts in WT animals. Pristane in the peritoneum is known to cause inflammation and damage (reviewed in (51)), and rapid desensitization of A₁R observed immediately after pristane injection is probably due to elevated levels of adenosine. The role of A_2AR in regulating lymphocyte depletion is controversial (38). It has been shown that elevations of cAMP following A_2AR activation can cause lymphocyte apoptosis (52, 53, 54). There are also studies showing the opposite – activation of A_2AR can protect lymphocytes from cell death (55, 56) We have previously published that the activation of A₁R and A_2AR are interdependent: the fast depletion of A₁R in the presence of elevated adenosine removes its anti‐apoptotic protection, and by reduction of G_i activity enables an early lymphotoxic effect by elevation of cAMP (25, 57, 58) it is possible that differences in A₁R expression or activation change A_2AR activity. In this study, we found that the effect of A₁R depletion is transient, and after 48 h lymphocyte counts begin to recover. We believe that the long‐term suppression of immunity that was observed is probably mediated by the elevated A_2AR (59), which is induced by A₁R, peaks at 48 h and remains high in WT animals even 10 days after pristane injection. The significant phenomenon of early A₁R stimulation following by up‐regulation of A_2AR was previously shown by our group (60). These findings also correspond with another work showing that following adjuvant‐induced arthritis, adenosine concentration in plasma remains high for weeks (61). In this study, pristane injection failed to up‐regulate A_2AR in A₁R‐KO mice, and the mRNA level of A_2AR was also lower in A₁R‐KO at T = 0. Hence, starting with low A_2AR levels leads to stronger lymphocyte reactivity in A₁R‐depleted animals and severe autoimmunity disease. A recent study shows a clear association between elevated cfDNA levels and autoimmunity (62).

In the T1DM model, cfDNA levels in A₁R‐KO mice were elevated compared to basal levels of WT mice. In both models, the higher levels of cfDNA were in accordance with the proportion of animals’ sickness. In the lupus model, we followed the levels of cfDNA in detail during the first 10 days after pristane injection. On day 10, A_2AR levels were low in A₁R‐KO mice compared to WT mice while cfDNA levels were significantly higher than the WT mice, which might contribute to the severe development of the disease in this group (63).

Adenosine receptors are expressed on a wide variety of immune cells (19). Several studies have shown that A_2AR has a role in inhibits TNF‐α secretion by macrophages (64, 65, 66) as well as an important role in alteration of macrophages from M1 macrophages to M2 macrophages (67, 68) which also contribute to autoimmunity development (69). All four adenosine receptors have been previously described in neutrophils (70). A₁R stimulation enhances neutrophils’ adherence to endothelium, chemotaxis (19) and their activity (71), while A_2AR inhibits neutrophil trafficking and effector functions such as oxidative burst, inflammatory mediator production and granule release (reviewed in (20, 72)).

In conclusion, adenosine initiates diverse cellular responses to prevent excessive inflammation and restoring immune homeostasis. Our data from the lupus model and T1DM suggest that A₁ and A_2A receptors have a protective role in autoimmunity development. The acute elimination of lymphocytes and reduction of DNA release due to NETosis depends upon A₁R desensitization and long‐term suppression maintained by A₁R‐dependent elevation of A_2AR. We believe that, based on these findings, severe traumatic events trigger a protective mechanism conducted by adenosine to reduce reaction against self‐antigens.

CONFLICT OF INTEREST

The authors have declared that no competing interests exist.

AUTHOR CONTRIBUTIONS

R.R. performed all experiments and, with the help of O.N., prepared the manuscript. J.M. was responsible for all hematological measurements and also helped in preparing the manuscript. Y.S.H. and C.C. helped in designing the study and acted as medical advisors. A.D. conceptualized the experiments, supervised all members of the research team, and helped in manuscript preparation. All authors contributed, in part, to the writing and editing of the final version.

DATA AVAILABILITY STATEMENT

The authors confirm that the data supporting the findings of this study are available within the article or available from the corresponding author (R. R.) upon reasonable request.