Foxp3⁺ regulatory T (Treg) cells play a pivotal role in the phenomenon of self-tolerance. In humans, Foxp3 mutations result in immunodysregulation, polyendocrinopathy, and enteropathy, with X-linked inheritance (called IPEX syndrome). Newborn boys with IPEX syndrome have severe enteropathy, eczema, type I diabetes, thyroiditis, hemolytic anemia, and thrombocytopenia; and they die within the first years of life if left untreated (1, 2). In the mouse model, Foxp3-deficient scurfy (SF) mice develop a lethal autoimmune disease which closely resembles the IPEX syndrome (3, 4). SF mice develop early-onset dermatitis, progressive multiorgan inflammation, and early death within the first month of life due to a lymphoproliferative syndrome. This lethal lymphoproliferative syndrome is predominately mediated by CD4⁺ T cells in humans and mice (5, 6). Consequently, the SF mouse is a valuable model for studying novel therapies for human IPEX syndrome and other autoimmune diseases associated with Treg deficiency. These include IPEX-like syndromes induced by mutations or deficiency in Itchy E3 ubiquitin protein ligase (ITCH), the α-chain of the IL-2 receptor (CD25), signal transducer and activator of transcription 5b, STAT1, or cytotoxic T-lymphocyte-associated protein 4 (7, 8).

High levels of the adenosine A_2A receptor are found in the brain, thymus, and spleen, as well as in circulating platelets and leukocytes (9). On the cell membrane of murine T lymphocytes, the adenosine A_2A receptor is highly expressed and is increased by T-cell receptor (TCR) stimulation (10, 11). In humans, the A_2A receptor is more highly expressed in CD4⁺ compared to CD8⁺ T cells (12). Moreover, numerous studies have highlighted the anti-inflammatory role of the adenosine A_2A receptor (13, 14). There have been observations of anti-inflammatory effects of A_2A receptor agonists in vivo and, conversely, enhanced inflammation in A_2A receptor knockout mice (14). However, the function of adenosine A_2A receptor in the development and control of autoimmune diseases remains unclear.

Recently, probiotics have emerged as relatively safe and inexpensive treatments for a number of gastrointestinal conditions. Lactobacillus reuteri strain DSM 17938 (L. reuteri) is a probiotic originally isolated from a Peruvian mother’s breast milk (15). This probiotic has been shown to prevent necrotizing enterocolitis (NEC) in newborn animals (16, 17) by inhibiting the toll-like receptor 4-mediated NF-κB pathway, facilitating the induction of immune-modulating Foxp3⁺ Tregs, and lowering the number of pro-inflammatory effector-memory T-cells in the intestinal mucosa. In humans, L. reuteri has been shown to reduce the severity of acute infant diarrhea (18–20), to prevent NEC in premature infants (21–23), and to decrease crying time in infants with colic (24, 25).

In addition, our recent studies demonstrated that L. reuteri significantly prolongs the survival rate of the SF mouse (from less than 30 days to greater than 4 months of age) by suppression of inflammatory T cells (mainly T_H1 and T_H2) extensively activated in multiple organs of SF mice (7). Mechanistically, L. reuteri modulates the abnormal microbial communities associated with these diseases, stimulating the production of bioactive metabolites involved in immune modulation. We observed that inosine, a downstream metabolite of adenosine, was decreased in the plasma of SF mice compared to wild-type (WT) mice, but was increased by oral administration of L. reuteri to SF mice. Oral administration of inosine by itself prolonged the survival and decreased autoimmunity of SF mice. Inosine was found to be a critical effector molecule of L. reuteri treatment, altering T_H1/T_H2 cell differentiation by activating A_2A receptors, predominately expressed on T cells. Blocking A_2A receptors by an A_2A antagonist reversed the anti-inflammatory effects of both inosine and L. reuteri, indicating that A_2A receptor appears to play a critical role in the beneficial effects of L. reuteri in the SF model (7).

In this study, we produced SF mice with genetically deleted adenosine A_2A receptor (SF⋅A2A-/-) to conclusively provide evidence of a central role of A_2A receptor in the actions of L. reuteri. We demonstrate that A_2A receptor gene deletion in SF mice did not accentuate the development of disease, but prevented the inhibitory effects of L. reuteri on autoimmunity. Our study highlights the A_2A receptor as a key mediator of the immunomodulatory mechanism of this probiotic.

This study demonstrated a central role of the adenosine A_2A receptor in mediating the protection of probiotic L. reuteri against inflammation in the Treg-deficient SF mouse (a model of human IPEX syndrome), evidenced by the observation that SF mice with an A_2A receptor deletion continued to have systemic inflammation which was unresponsive to L. reuteri treatment.

It is well known that the lethal lymphoproliferative syndrome characterizing SF mice is predominately mediated by T_H1 and T_H2 cell-induced pathology (27, 28). The key to Treg suppression of T effector cells (T_H1/T_H2/T_H17) is an interaction between adenosine produced by Tregs (mediated by a CD39–CD73 pathway) and the A_2A receptor expressed on nearby T effector cells (31). Lymphocytes predominately express A_2A receptors (10–12, 32, 33). However, during Treg deficiency in SF mice or human IPEX syndrome, T_H1 and T_H2 cells lose their regulation by adenosine A_2A-mediated signaling, resulting in T_H1 and T_H2 cell-induced pathology. Studies by Csoka et al. showed that an agonist of A_2A receptors inhibited the proliferation and effector functions of CD4⁺ T cells isolated from WT mice but failed to block these of cells obtained from A_2A knockout mice (33), indicating that the activated adenosine A_2A receptor plays a critical role in the suppression of T_H1 and T_H2 cells.

Our previous study demonstrated that Treg deficiency induces gut microbial dysbiosis dynamically over the first 22 days of life, an effect which could be reprogrammed by oral administration of L. reuteri. L. reuteri suppressed T_H1 and T_H2 cells in SF mice, as evidenced by lower circulating levels of IFN-γ (T_H1) and IL-4 (T_H2) and reduced numbers of IFN-γ and IL-4-expressing lymphocytes in spleen and mesenteric lymph nodes of SF mice. Metabolites produced by L. reuteri or L. reuteri-modulated bacteria are known to promote or suppress immune cell function (34–36). We discovered that the purine metabolite inosine, a metabolite of adenosine, is severely decreased in SF mice, while increased after oral administration of L. reuteri (7). Inosine has been proved to be a functional agonist of the A_2A receptor which has an anti-inflammatory effect (37–43). Our previous experiments by using adenosine receptor knockout mice to study the suppression of inosine on naïve CD4⁺ T cell differentiation into T_H1 and T_H2 in vitro strongly suggested that the effects of inosine are dependent on the A_2A receptor on T cells (7). In addition, an in vivo study showed that an A_2A receptor antagonist blocks the anti-inflammatory effects of both inosine and (L. reuteri DSM 17938) on T_H1 and T_H2 suppression and multiorgan lymphocyte infiltration in SF mice (7). In summary, the A_2A receptor mediates the beneficial biological effects of L. reuteri and inosine in SF mice. In this study, we further confirmed a critical role of A_2A receptor-mediated effects by genetic deletion of A_2A in SF mice (SF⋅A2A-/- mice).

Mechanistically, how L. reuteri results in increased serum level of inosine is not fully understood. When we compared L. reuteri cultures to MRS broth (culture media without L. reuteri) after 16 h of anaerobic growth, L. reuteri did not generate significant amounts of purines or inosine in culture. Our previous studies indicated that enterally feeding L. reuteri is associated with recovery of the plasma levels of inosine and hypoxanthine to levels similar to WT, at the same level that inosine levels decreased in the stool of these mice (7). We hypothesized that, most likely, L. reuteri promotes inosine absorption in the intestine by improving overall gut health through multiple mechanisms (for example, by improving villus length) and/or by modulating the gut microbial community. We measured the small intestinal villi in SF mice compare with SF mice after oral feeding L. reuteri and showed that orally feeding L. reuteri improves the length of villi and depth of crypts. Furthermore, an increased expression of equilibrative nucleoside transporter transporters after L. reuteri feeding was found, which could contribute to produce improved absorption. The best method to confirm enhanced absorption would be to orally feed labeled inosine after administration of L. reuteri and quantify the labeled inosine in the circulation. However, the labeling approach for small molecules like inosine is much more difficult than for amino acid or proteins. In the meantime, we could not rule out that in vivo the gut environment could activate the enzymes such as adenosine deaminase (ADA) and 5′-nucleotidase generated by L. reuteri to produce inosine. But it is difficult to distinguish the ADA activity in the intestinal tissue lysates from the activity of L. reuteri or other microbes, because ADA activity is very high in the intestine (44). The direct links between L. reuteri and the metabolites required further exploration.

We also noticed that A_2A receptor appears to be expressed in other organs besides lymphocytes (11, 45). In liver, the A_2A receptor is expressed in Kupffer cells, hepatocytes, and hepatic stellate cells (46–48). Some studies suggested that the A_2A receptor plays a role not only in regulating inflammation but also in maintaining liver function in general (39). Previous studies also revealed that it is more highly expressed in spleen, lymph nodes, liver, and lung than that in the small intestine or adrenal gland, supporting a functional role of this receptor in the regulation of the immune response in peripheral lymphoid tissues (11). It has been reported that A_2A receptor activation confers tissue protection in peripheral organs (49, 50). While the mechanism of L. reuteri in regulating inflammation in SF mice clearly involves T cell modulation, we cannot rule out that A_2A receptor expression in these organs may also contribute to the beneficial effects of L. reuteri in SF mice. Therefore, A_2A receptor expression on both immune cells and other cells and their interaction may determine the overall impact of A_2A receptor deletion on beneficial effects of L. reuteri.

The role of the T cell and its expression of A_2A modulated by L. reuteri or highly related metabolites such as inosine could be further studied by using a T cell knockout mouse model by adoptive transfer of CD4⁺ T cells isolated from WT, SF, A2A-/-, or SF⋅A2A-/- with/without L. reuteri or inosine treatment, which is currently under investigation.

In summary, our study demonstrates that adenosine A_2A receptor deletion does not inhibit the development of autoimmune disease in the SF mouse. However, adenosine A_2A receptor deletion reverses the inhibition of L. reuteri on autoimmunity induced by Treg-deficiency in SF mice. Our results support the concept that activated adenosine A_2A receptors are linked to L. reuteri effects in vivo. They also suggest that the activated A_2A receptor by L. reuteri or other agonists may represent a useful therapeutic strategy for preventing lethal outcomes in Foxp3 deficency- or dysfunction-induced autoimmune diseases.

This study was carried out in accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals (NIH) and The Institutional Animal Care and Use Committee (IACUC). The protocol was approved by the IACUC (Protocol number: AWC-14-056 and AWC-17-0045).

BH, YL, and JR conceived and designed the experiments. BH, TH, and YL performed all experiments and analyzed the data. BH, YL, DT, and JR wrote the paper and edited the manuscript. All authors read and approved the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.