The β2-adrenergic receptor controls inflammation by driving rapid IL-10 secretion

Abstract

The mammalian nervous system communicates important information about the environment to the immune system, but the underlying mechanisms are largely unknown. Secondary lymphoid organs are highly innervated by sympathetic neurons that secrete norepinephrine (NE) as the primary neurotransmitter. Immune cells express adrenergic receptors, enabling the sympathetic nervous system to directly control immune function. NE is a potent immunosuppressive factor and markedly inhibits TNF-α secretion from innate cells in response to lipopolysaccharide (LPS). In this study, we demonstrate that NE blocks the secretion of a variety of proinflammatory cytokines by rapidly inducing IL-10 secretion from innate cells in response to multiple Toll-like receptor (TLR) signals. NE mediated these effects exclusively through the β2- adrenergic receptor (ADRB2). Consequently, Adrb2^−/− animals were more susceptible to L. monocytogenes infection and to intestinal inflammation in a dextran sodium sulfate (DSS) model of colitis. Further, Adrb2−/− animals rapidly succumbed to endotoxemia in response to a sub-lethal LPS challenge and exhibited elevated serum levels of TNF-α and reduced IL-10. LPS-mediated lethality in WT animals was rescued by administering a β2-specific agonist and in Adrb2−/− animals by exogenous IL-10. These findings reveal a critical role for ADRB2 signaling in controlling inflammation through the rapid induction of IL-10. Our findings provide a fundamental insight into how the sympathetic nervous system controls a critical facet of immune function through ADRB2 signaling.

INTRODUCTION

Inflammation is a tightly controlled process that ensures proper localization of immune cells, release of pro- and anti-inflammatory mediators, clearance of dead cells, and containment of the pathogen ^1,2 Intrinsic to the immune system are a series of checks and balances which are in place to prevent auto-reactivity and hyper inflammation. In addition, the immune system is also under direct control by the central nervous system. Early studies demonstrated that both primary and secondary lymphoid tissue s are innervated by post-ganglionic sympathetic neurons that secrete NE as the primary neurotransmitter ^{3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13} Likewise, immune cells express adrenergic receptors, which bind both epinephrine (E) and NE, giving the immune system the ability to directly respond to signals from the autonomic nervous system. E is primarily secreted by the chromaffin cells within the adrenal gland and exerting systemic effects, whereas NE is secreted predominantly by sympathetic nerves that terminate locally within peripheral organs. At rest, adrenergic neurons maintain the autonomic activity of the organs they regulate, such as heart rate, blood pressure, and gut peristalsis. Given the intimate association of sympathetic neurons and immune cells within lymphoid organs, NE represents the primary neurotransmitter that can immediately impact immune function ^{14, 15}

While signaling through pattern recognition receptors (PRRs) promotes inflammatory cytokine secretion from antigen presenting cells, neurons themselves express various TLRs, enabling them to respond directly to certain pathogen-associated molecular patterns (PAMPs) ^{16, 17, 18, 19, 20, 21,22} Consequently, sympathetic neurons secrete NE in response to pathogenic organisms (reviewed in ¹³). Both viral and bacterial infections elicit bursts of NE secretion from sympathetic neurons, and PAMPs such as lipopolysaccharide (LPS) drive NE released rapidly upon exposure ^{23, 24} As such, NE carries the potential to regulate a variety of immune functions both systemically and locally. Perhaps most prominent are the anti-inflammatory effects of NE exerted on both innate and adaptive cells. Adrenergic receptor signaling on T cells has been shown to regulate their trafficking to peripheral tissues and through secondary lymphoid organs ^{25, 26, 27} Recently we ²⁸ and others ^{29, 30} found that NE inhibited inflammatory cytokine secretion and cytolytic activity in both mouse and human effector CD8⁺ T cells. NE also indirectly limits the magnitude of CD8⁺ T cell priming by suppressing cross-presentation in dendritic cells ³¹. In macrophages, NE suppresses TLR4-dependent secretion of TNF-α, resulting in reduced levels of inflammation and sepsis in response to LPS ³², and these effects are also observed in populations of dendritic cells ^31,33,34. One mechanism that has been proposed for NE-mediated suppression of LPS-induced TNF-α secretion involves the adaptor molecule β-arrestin, which is post-translationally regulated by adrenergic signaling and blocks NFκB activation ³⁵ Flowever, it remains unclear how adrenergic signaling broadly impacts inflammatory pathways and what mechanisms regulate general immune suppression mediated by NE. In this study, we found that NE globally suppressed a variety TLR- induced pro-inflammatory cytokines while rapidly inducing the expression of the ant-inflammatory cytokine IL-10. NE promoted these activities rapidly and specifically through the β2-adrenergic receptor (ADRB2). In experimental models of infection, DSS- induced colitis, and sepsis, Adrb2^−/− mice were unable to control inflammation and were highly susceptible to LPS-mediated endotoxemia. Their inability to control inflammation correlated with increased serum TNF-α and decreased IL-10. Finally, Adrb2-deficient mice could be rescued from lethal endotoxemia with a single dose of rmlL-10. Further, administration of the long-acting β2-agonist salmeterol enhanced IL-10 expression in response to LPS and improved overall survival in WT animals. This study highlights an unprecedented role for the sympathetic nervous system in controlling inflammation and uncovers a critical pathway regulating the secretion of the anti-inflammatory cytokine IL- 10.

RESULTS

NE globally suppresses TLR-induced pro-inflammatory cytokine secretion.

Previous studies have demonstrated that ADRB2 signaling suppressed LPS-induced TNF-α secretion ^36,37 via stabilization of IκB and blockade of NF-κΒ by β-arrestin ³⁵ We wished to determine if this effect was exclusive to a TLR4 ligand and whether expression of other pro-inflammatory cytokines was regulated by NE. Bone marrow derived macrophages (BM-DM) were incubated with the TLR ligands Pam3CSK4, Polyl:C, LPS, R837 (Fig. 1A), and CpG (Supplementary Figs. 1 and 2) in the absence or presence of NE. As reported previously³³ NE potently suppressed TNF-α secretion driven by LPS activation. Flowever, this effect was not exclusive to a TLR4 agonist as NE also suppressed TNF-α secretion from every TLR agonist tested (Fig. 1A). Moreover, NE suppressed other pro-inflammatory cytokines including IL-12, IL-6 (Fig. 1B), and KC/GRO (Supplemental Fig. 1) regardless of which TLR pathway was engaged. Of note, we found that IL-6 secretion was slightly enhanced by NE at 2 hrs post-stimulation, but markedly suppressed by NE at all subsequent time intervals. While these TLR agonists do not efficiently drive inflammasome activation, the low levels of IL- 1β that were secreted in these assays were also suppressed by NE (Supplemental Fig. 1). Finally, NE suppressed TNF-α, IL-12 and IL-6 secretion in a dose dependent manner at concentrations as low as 0.5 nM (Supplemental Fig. 2).

Figure 1. NE suppresses TLR-induced pro-inflammatory cytokine secretion in bone marrow-derived macrophages.

(A) BM-DMs from Balb/cJ mice were stimulated with one of the following TLR agonists: 100 ng/mL Pam3CSK4, 10 μg/mL poly l:C, 100 ng/mL LPS, or 1 ng/mL R837 in the absence or presence of 5 μΜ NE for 2, 8 and 48 hours. Error bars represent the standard error of the mean for n=2 mice. (B) BM-DMs from C57BI/6J mouse were stimulated in the presence or absence of 5 μΜ NE and 1 μΜ CpG for 2–48 hours. Cytokines in the supernatant were quantified using Mouse Pro- inflammatory 7-Plex Tissue Culture Kit from Mesoscale Discovery.

As both α and β class adrenergic receptors (AR) can be expressed by innate immune cells, we sought to determine whether the general suppressive activity of NE was mediated by a single or multiple ARs. We utilized receptor antagonists to distinguish which ARs were involved in suppression of pro-inflammatory cytokines. As above, NE suppressed CpG-induced TNF-α secretion, which was reversed by the pan βAR blocker nadolol, but not by the pan αAR blocker phentolamine (Fig. 2A). NE- mediated suppression was also blocked by the ADRB2-specific inhibitor ICI118,551, but not the βΙ-AR-specific antagonist atenolol, indicating a selective role for the ADRB2 in NE-mediated suppression. In parallel, BM-DMs from Adrb2^−/− animals were completely resistant to NE-mediated suppression of pro-inflammatory cytokine secretion driven by all TLR ligands tested (Figure 2B), indicating that the ADRB2 is the sole receptor involved in this pathway. Bone marrow-derived dendritic cells were also suppressed by NE in a similar manner (Supplemental Fig. 3). Finally, the β2-specific agonist albuterol suppressed CpG-induced TNF-α secretion in murine BM-DMs and reduced TNF-α production in LPS-treated human whole blood, suggesting an evolutionary conserved pathway between mice and humans (Figure 2C and D).

Figure 2. NE suppresses TLR-induced pro-inflammatory cytokine secretion through the ADRB2.

(A) BM-DMs from C57BI/6J mouse were stimulated with one of the following adrenergic receptor antagonists: 1 μΜ phentolamine (aAR antagonist), 1 μΜ nadolol (PAR antagonist), 100 nM atenolol (P1AR antagonist), or 100 nM ICI118,551 (P2AR antagonist) in the presence or absence of 5 μΜ NE and 1 μΜ CpG for 24 hours. TNF-α in the supernatants was quantified by ELISA. (B) WT and Adrb2^−/− BM-DMs were stimulated with different TLR ligands in the presence or absence of norepinephrine for 24 hours. Supernatants were harvested and TNF-α in the supernatant was quantified by ELISA. Error bars represent the standard error of the mean for n=3 mice per strain. (C) BM-DMs were stimulated with 1 μΜ CpG and increasing concentrations of albuterol for 2, 8 and 48 hours. Error bars represent the standard error of the mean for n=3 mice per strain. (D) Whole blood from healthy adult donors were stimulated with 10 μg/mL LPS in the presence or absence of 50 nM albuterol for 4 hours. Plasma TNF-α was assessed by ELISA. The data represents 6 donors. Two-tailed paired t-test was conducted for (D), p<0.05.

ADRB2 signaling selectively drives rapid IL-10 expression.

As opposed to TNF-α, TLR stimulation induces expression of IL-10 relatively late during activation as a mechanism to limit pro-inflammatory cytokine accumulation over the course of inflammation ³⁹ Although LPS is a rather weak inducer of IL-10 transcription compared to other TLR pathways, IL-10 expression is critical to suppress inflammatory pathways ^40,41 and to dampen TLR signaling by suppressing the expression and activation of downstream signaling intermediates ^42,43 Recent reports have demonstrated that adrenergic signaling can enhance IL-10 expression in dendritic cells ³⁴ and macrophages ³² However, it was not clear whether IL-10 induction was a general property of NE signaling in the context of all TLR signals and whether this could explain the anti-inflammatory properties of NE. To address this, IL-10 secretion was measured from BM-DMs cultured with CpG in the absence or presence of NE. CpG induced IL-10 secretion, which peaked at 8 hrs post-stimulation and remained steady to 24 hrs (Fig. 3A). However, in contrast to TNF-α and other inflammatory cytokines, NE markedly enhanced IL-10 secretion in response to CpG at early time points, including 2 hrs poststimulation. This effect synergized with TLR signaling as IL-10 secretion was nominal in response to NE alone (Fig. 3A), and this induction required expression of the ADRB2 (Fig. 3B). Moreover, NE induced early IL-10 secretion in response to all TLR ligands tested, demonstrating a general role for adrenergic signaling in driving acute IL-10 secretion in macrophages (Fig. 3C), which was paralleled by suppressed Tnf and increased Il10 mRNA transcripts (Supplemental Fig. 4). Finally, we found that NE also suppressed TNF-α and drove acute IL-10 secretion from human monocyte-derived macrophages (Fig. 3D). Collectively, these data demonstrate that signaling via the ADRB2 converts a potent inflammatory pathway into an anti-inflammatory pathway by suppressing pro-inflammatory cytokines and inducing early secretion of IL-10. The acute nature of this response provides a rapid control of inflammation by the neuroendocrine system. Given that this effect synergized with all TLR pathways tested, it is likely that adrenergic signaling is a general anti-inflammatory pathway that potently modulates the magnitude of inflammation.

Figure 3. Engagement of adrenergic receptors induce rapid IL-10 production.

(A) BM-DMs from C57BI/6J mouse were stimulated in the presence or absence of 5 μΜ NE and 1 μΜ CpG for 2–48 hours. (B) Adrb2- sufficient and -deficient BM-DMs were stimulated with 100 ng/mL LPS in the presence or absence of norepinephrine for 24 hours. Supernatants were harvested and IL-10 in the supernatant was quantified using Mouse Pro-inflammatory 7-Plex Tissue Culture Kit from Mesoscale Discovery. Error bars represent the standard error of the mean for n=3 mice per strain. (C) BM-DMs were stimulated with one of the following TLR agonists: 100 ng/mL Pam3CSK4, 10 pg/mL poly l:C, 100 ng/mL LPS, 1 pg/mL R837, or 1 μΜ CpG for 2 hours in the absence (black bars) or presence (red bars) of 5 μΜ NE. The cytokines released to the supernatant were quantified using Mouse Pro-inflammatory 7-Plex Tissue Culture Kit from Mesoscale Discovery. (D) Human monocyte-derived macrophages were stimulated with 10 pg/mL LPS in the presence or absence of 5 μΜ NE for 2 hours. Secreted TNFa and IL-10 were quantified by ELISA. The data represents 4 donors. **p<0.001 by Bonferroni post-hoc test following two-way ANOVA (B) and p<0.05 by two- tailed paired t-test (D).

TNF-α is one of the main cytokines that mediates many of the hallmark inflammatory processes in cells and tissues. A recent study demonstrated that NE suppressed TNF-α secretion from macrophages by inhibiting NFκB activation via β- arrestin ³⁵ Whether IL-10 was involved in this process was not investigated. To address the general role of IL-10 in NE-mediated pro-inflammatory cytokine secretion, we measured TNF-α, IL-6, and IL-12 secretion in Il-10^−/− macrophages (Fig. 4). BM-DMs from WT and Il-10^−/− mice were stimulated with either CpG (Fig. 4A-C) or LPS (Fig. 4D- F) in the absence or presence of NE. In general, both CpG and LPS induced secretion of all three cytokines, which plateaued at approximately 8 hrs post-stimulation. However, we observed clear differences in 1) the kinetics of cytokine expression between CpG and LPS stimulation, 2) the suppression of each cytokine by NE, and 3) the requirement for IL-10 in NE-mediated cytokine suppression. For example, TNF-a was induced rapidly by both CpG and LPS and was inhibited by NE in Il-10^−/−macrophages at early timepoints (2–4 hrs, Figs. 4A and Figs. D). However, TNF-α secretion was increased in Il-10^−/− cells at later timepoints in response to CpG + NE, but not with LPS + NE, which suggests differential sensitivity to IL-10 depending upon the specific TLR pathway. In contrast, early IL-6 secretion was markedly increased in Il-10^−/−macrophages in response to both CpG and LPS, yet NE-mediated suppression of IL-6 was completely dependent upon IL-10 at all timepoints except for 2 hrs (Fig. 4B and E). Finally, although IL-12 expression at 2 hrs was quite low, it failed to be suppressed by NE at the 2 hrs, but not the later timepoints in Il-10^−/− cells (Fig. 4C and F). Thus, in the context of genetic ablation, IL-10 was dispensable for NE-mediated TNF-α suppression at early timepoints, while NE-mediated suppression of IL-6 and IL-12 was variably dependent on IL-10 at different time points of activation.

Figure 4. NE-mediate regulation of cytokine expression in II–10−/− macrophages.

.WT and Il10^−/− BM-DMs were stimulated with 1 μΜ CpG (A-C) or 100 ng/mL LPS (D-F) in the presence or absence of 5 μΜ NE for 2, 4, 8, and 24 hours. Cytokines in the supernatant were quantified using Mouse Pro-inflammatory 7-Plex Tissue Culture Kit from Mesoscale Discovery. Bar graphs on the left of each panel are data from the 2 hr timpoint that are included in the kinetic data displayed in the line graphs on the left of each panel. Error bars represent the standard error of the mean for n=2–3 mice per strain. p<0.05 by two-tailed paired t-test within each strain.

As IL-10 is important for the maturation and appropriate development of many immune cell types, it is possible that Il-10−/− macrophages engage alternative pathways to prevent abortive development. To address the direct role of autocrine IL-10 effects on NE-mediated cytokine suppression, we utilized an anti-IL-10 receptor (a-IL-10R) blocking antibody (Fig. 5). Here, we compared the expression of cytokines in response to CpG and Pam3CSK4. TNF-α was potently suppressed by NE in response to both CpG and Pam3CSK4 at both 2 and 8 hrs post-stimulation (Fig. 5A and D). In contrast to results obtained with Il-10−/−cells, we found that NE-mediated suppression of TNF-α was reversed by a-IL-10R treatment in the context of CpG activation at 8 hrs. Suppression of TNF-α was minimally sensitive to a-IL-10R blockade when activated by Pam3CSK4, indicating a selective requirement for IL-10 in NE-mediated suppression. However, NE- mediated suppression of both IL-6 and IL-12 required autocrine IL-10R activity, particularly at the 8 hr timepoint (Fig. 5B-C and E-F). These results reveal the complex nature of ADRB2 signaling and the differential sensitivity of each TLR pathway to NE- mediated suppression. Nonetheless, we observe a consistent suppression of pro- inflammatory cytokine secretion by NE that is contextually and temporally dependent upon IL-10.

Figure 5. The role of the IL-10R in NE-mediated suppression of the pro-inflammatory cytokine secretion.

BM-DMs from Balb/cJ mouse were pre-treated with 10 pg/mL anti-CD210 (IL-10R) antibody or isotype control for 30 minutes. BM-DMs were stimulated with 1 μΜ CpG (A-C) or 100 ng/mL Pam3CSK4 (D-F) in the absence or presence of 5 μΜ NE for 2 and 8 hours. Cytokines were quantified with Mouse Pro- inflammatory 7-Plex Tissue Culture Kit from Mesoscale Discovery. p<0.05 by Tukey’s post-hoc test following ordinary one-way ANOVA across treatment groups.

ADRB2 signaling controls inflammation in vivo.

Inflammation is balanced by processes that sterilize infected tissue and by the effects of those processes on collateral tissue damage. In the context of infection, it is possible that ADRB2 signaling could either promote or suppress an appropriate sterilizing immune response. We tested the response to a systemic L. monocytogenes (LM) infection in WT and Adrb2^−/− mice. Upon lethal exposure to LM, Adrb2-deficient animals succumbed to the infection more rapidly than Adrb2-sufficient animals (Figure 6A). In a sub-lethal challenge model, we observed ~5-fold increase in bacterial burden at day 3 within the spleens of Adrb2^−/− animals, which was not statistically significant when compared to WT mice (Supplemental Fig. 5A and B). However, we observed enhanced splenomegaly, but not hepatomegaly (data not shown) in Adrb2-deficient animals at day 7 post-infection, indicating hyperinflammation within the spleen at this timepoint (Fig. 6B). We observed a significant expansion of total splenocytes in Adrb2^−/− mice, which were comprised of CD11b⁺ innate cells, CD3⁺ T cells, but no change in CD19⁺ B cells compared to infected WT animals (Supplemental Fig. 5C). Finally, the overall inflammation in WT animals could be further reduced by treatment with the long-acting β2-agonist salmeterol (Fig. 6C), indicating a specific role for the ADRB2 in regulating inflammation during a live intracellular pathogen infection.

Figure 6. Adrb2 signaling controls innate responses to bacterial infection.

(A) WT and Adrb2^−/− mice were challenged with a lethal dose of LM (LM-OVA, 277K CFU/mouse) intravenously (i.v.) and the survival of the animals were monitored for 7 days. (n=10–11 mice per strain). (B) WT and Adrb2^−/− mice were challenged with sub- lethal dose of LM-OVA i.v. and spleen weights were assessed during infection as a proxy for inflammation (n=7–16 mice per strain). (C) WT and Adrb2^−/− mice were challenged with sub-lethal LM-OVA and were administered 40 pg Salmeterol at days 0, 1 and 2 of infection intraperitoneally. Spleen weights were assessed day 7 postinfection (n=12–16 mice per condition). p=0.0422 by log-rank test (A), p<0.05 by Bonferroni post-hoc test following two-way ANOVA (B), p<0.05 by Tukey’s post-hoc test following two-way ANOVA (C).

ADRB2 signaling synergized with TLR activation to drive early secretion of IL-10 from innate cells (Fig. 3), suggesting an important role for the ADRB2 in controlling inflammation via IL-10. In vivo, IL-10 and IL-10R deficiencies lead to a variety of auto- inflammatory processes, of which colitis is the most pronounced phenotype ⁴⁴ As such, we tested whether the ADRB2 was involved in inflammation by promoting colitis with oral consumption of dextran sodium sulfate (DSS) ⁴⁵ Both WT and Adrb2^−/− failed to gain weight during the early course of DSS treatment, and Adrb2^−/− lost significantly more weight than WT mice by days 6–7 (Fig. 7A). DSS-induced weight loss correlated with 10% shortened colon lengths in Adrb2^−/− animals (Fig. 7B), which was not significant when compared to WT colons at this time point of treatment. However, there was a direct correlation between colon length and histopathology scores (Fig. 7C), however, there were no statistical differences in pathology scores when comparing WT to Adrb2^−/− colons (data not shown).

Figure 7. Adrb2-deficient animals are susceptible to DSS colitis.

WT and Adrb2^−/− mice were given either water or 3% DSS diluted in their drinking water for 7 days, and body weights were measured for each animal during the course of the experiment. (A) The percent of starting body was calculated through d7, at which time animals were sacrificed to measure colon length (B) and histopathology. (C) Linear regression analysis of colon length versus histopathology scoring of colon inflammation. (D) Representative histopathology images of colons from two separate mice from either the WT or Adrb2^−/− groups. n=7–8 mice per strain. p<0.05 by Bonferroni post-hoc test following two-way ANOVA.

We extended our analysis of inflammation by employing an in vivo model of sepsis with systemic LPS administration. This classic model of sepsis drives successive secretion of IL-1 β, TNF-α, and IL-6, and the lethality can be altered by titrating the dose of LPS. Here, Adrb2^−/− animals were exquisitely sensitive to an otherwise sublethal dose of LPS, with approximately 80% of animals succumbing to endotoxemia by 48 hrs following 300 μg of LPS administration (Fig. 8A). At this dose, there was a corresponding increase in TNF-α and decreased concentration of systemic IL-10 at 2 hrs in the serum of Adrb2^−/− mice compared to WT controls (Fig. 8B). These findings support a general role for the ADRB2 in suppressing inflammation via induction of IL-10, as we observed with in vitro culture of BM-DMs. However, because both immune and non-immune cells can contribute to systemic cytokine responses to LPS in vivo, we employed the use of the conditional Adrb2^fx/fx mouse model ⁴⁶ to determine the contribution of the ADRB2 on innate immune cells during endotoxemia. Adrb2^fx/fx mice were crossed to the LysM-Cre strain ⁴⁷ to generate animals in which the Adrb2 was deleted primarily in innate cells, including dendritic cells, neutrophils and macrophages. Similar to our findings in the germline Adrb2^−/− deletion, we found that selective deletion of the Adrb2 in innate cells was sufficient to promote lethal endotoxemia to an otherwise non-toxic dose of LPS (Fig. 8C).

Figure 8. Adrb2-deficient animals are susceptible to endotoxemia.

WT and Adrb2^−/− mice were challenged with 300 pg LPS i.p. (A) Survival of the animals was monitored for 7 days (n=18–20 mice), and (B) serum cytokines were assessed in separate cohorts of animals by ELISA (n=9–14 mice). (C) LysM-Cre⁺ WT, Adrb2^fx/+ and Adrb2^fx/fx mice were challenged with 250 pg LPS i.p. and the survival of the animals were monitored for 7 days (n=4–8 mice). p<0.05 by log-rank test (A and C), p<0.05 by Bonferroni post-hoc test following two-way ANOVA.

Adrenergic signaling regulates a host of physiological processes and is critical for modulating blood pressure, heart rate, and body temperature in both homeostasis and during physical crisis. Although the reduction in IL-10 levels could be correlated with increased susceptibility in the absence of the Adrb2, it was also possible that these animals were unable to regulate basic physiological responses to low doses of LPS, which would be independent of the effects of IL-10. To determine the role of IL-10 in the context of ADRB2 signaling deficiency, we first tested whether IL-10 treatment could influence the course of endotoxemia in Adrb2-deficient mice. As shown in Fig. 9A, a single dose of IL-10 (1 μg) given at the time of LPS administration significantly rescued Adrb2^−/− animals from lethal endotoxemia. In isolation, we found that signaling via the ADRB2 was not sufficient to drive significant IL-10 secretion from macrophages or dendritic cells in the absence of a TLR signal. We explored the possibility that in vivo activation of the ADRB2 with a pharmacological β2-specific agonist in combination with TLR stimulation would drive increased IL-10 secretion and perhaps alter the course of LPS-induced endotoxemia. To test this, we first titrated LPS to a dose that we found was 50% lethal in a cohort of WT Balb/c mice, which empirically was found to be 750 μg (data not shown). Cohorts of animals were then treated with LPS in the absence or presence of the long-acting β2-specific agonist salmeterol. Approximately 50% of animals succumbed to endotoxemia, which was significantly rescued by salmeterol (Fig. 9B). Activation of the ADRB2 by salmeterol led to a significant decrease in TNF-α and increase in IL-10 2 hrs following LPS administration (Fig. 9C). These results indicate that ADRB2 drives expression of IL-10, which may act to block the overall lethal inflammation driven by LPS. Indeed, we observed rapid and 100% lethality in animals treated with an IL-10R blocking antibody, which was not affected by co-treatment with salmeterol (Fig. 9B). Thus, as we found in isolated cultures of macrophages and dendritic cells, ADRB2 signaling synergizes with TLR activation to drive rapid IL-10 secretion, which is critical for survival from LPS-mediated endotoxemia.

Figure 9. ADRB2 signaling controls IL-10 production in vivo.

(A) WT and Adrb2^−/− mice were challenged i.p. with 200–250 μg LPS. A cohort of Adrb2^−/− were administered 1 μg recombinant murine IL-10 i.p. immediately after LPS challenge. The survival of the animals was monitored for 7 days (n=13–14 mice per condition) (B) WT Balb/c mice were challenged with 750 pg LPS, in the presence of 40 pg salmeterol (or vehicle control) and/or 200 pg anti CD210 (IL-10R) antibody (or isotype control). The survival of animals was monitored for 7 days. (n=17–20 mice per treatment group) (C) WT Balb/c mice were challenged with 300 pg LPS in the absence or presence of 40 pg salmeterol. Blood was harvested at indicated time points and serum cytokines were quantified by ELISA. (n=3–10 mice per time point). p<0.05 by log-rank test (A and B), p<0.05 by unpaired t-test (c).

DISCUSSION

The adrenergic system is perhaps one of the best characterized autonomic physiological regulators. As a component of the sympathetic nervous system, adrenergic signaling complements a variety of neuroendocrine effectors including cortisol and acetylcholine, which are known modulators of immune responses ^48,49,50 The immunosuppressive activity of adrenergic signaling has been demonstrated by the ability of epinephrine, NE, and β-agonists to markedly inhibit LPS-induced TNF-a secretion from macrophages. More recently, Grailer et. al. demonstrated that the β2- specific antagonist, ICI 118,551, could increase susceptibility to LPS-induced endotoxemia in mice ³², highlighting the potential requirement for the ADRB2 in immune suppression. In this study, we uncovered a critical role for the ADRB2 in converting a general inflammatory signal into an anti-inflammatory pathway by synergizing with TLR signaling to induce the rapid transcription and secretion of IL-10. Moreover, we found that the ADRB2 expressed predominantly by myeloid cells was critical for the susceptibility to endotoxemia, likely through induction of IL-10, which was significantly increased by treatment of animals with the long-acting β2-agonist salmeterol. ADRB2 activation alone marginally induced IL-10 mRNA expression without a concomitant increase in IL-10 secretion (data not shown). However, regardless of which TLR was engaged, ADRB2 signaling synergistically promoted high acute expression and secretion of IL-10.

The early induction of IL-10 correlated with a profound suppression of the inflammatory cytokines TNF-α IL-6, and IL-12 at later timepoints in vitro. While TNF-a secretion was impacted early during activation, suppression of IL-6 and IL-12 was delayed, which may indicate distinct pathways of regulation for these cytokines. Previous studies have demonstrated a role for β-arrestin2 in suppressing LPS-induced TNF-α section in response to NE by stabilizing IκΒα and blocking NFκB-dependent induction of TNF-α transcription ^35,51,52. In this situation, NE-mediated TNF-a suppression would be predicted to be independent of the autocrine suppressive actions of IL-10, which has been shown previously in a stress-induced reflex model of endotoxemia ⁵³ Indeed, we found that NE could suppress TNF-α secretion in Il-10^−/− macrophages in response to both CpG and LPS. Moreover, blocking the IL-10R in WT macrophages failed to reverse NE-mediated TNF-α suppression at the early 2 hr timepoint by Pam3CSK4 and only slightly in response to CpG. However, prolonged TNF-α secretion was restored at the 8 hr timepoint and beyond by IL-10R neutralization. Moreover, IL-12 and IL-6 secretion levels were fully restored by blocking the IL-10R in WT cells, indicating a critical role for IL-10 in NE-mediated suppression of these two cytokines. Thus, suppression of TNF-α by ADRB2 signaling may operate by dual pathways: one in which ADRB2 activation mobilizes β-arrestin2 to block early NFkB- dependent TNF-α transcription, and a second later inhibition of TNF-α secretion mediated by autocrine IL-10 signaling. This dual mechanism likely underlies the observed reciprocal control of TNF-α and IL-10 in vivo in response to LPS, in which case splanchnic nerve ablation increases overall TNF-α concentrations at the expense of early IL-10 secretion ⁵⁴ B-arrestin2 may also be required for the rapid induction of IL- 10 as β-arrestin2 has been shown to bind to and stabilize p³⁸ for activation by TLR4 signaling ⁵⁵ In this case, β-arrestin2 was required for TLR-dependent IL-10 expression even in the absence of ADRB2 engagement. Taken together, we propose a mechanism which ADRB2 signaling drives rapid IL-10 expression in the context of TLR activation, which can act in an autocrine fashion to suppress pro-inflammatory cytokine secretion. Additional mechanisms, potentially involving β-arrestin2, act early in ADRB2 signaling to block TNF-α. While β-arrestin2 may inhibit NFκB-dependent TNF-α transcription, this mechanism cannot explain the general suppressive effects of ADRB2 signaling, as IL- 10 transcription is also partially dependent on NFκB activation ^56,57.

While our in vitro observations suggested a robust immunosuppressive function for ADRB2 signaling, we wished to determine the magnitude of this pathway as it regulates inflammation in vivo. We used three different inflammatory models to determine how endogenous ADRB2 signaling broadly influenced inflammation. First, we tested the response to a live LM infection, as previous studies suggested that the ADRB1 may be required for controlling LM in the context of external cold stress ⁵⁸ Indeed, ADRB2-deficient animals were marginally more susceptible to a lethal inoculum LM than WT mice, which correlated to enhanced splenomegaly and an influx of both innate cells and T cells during the course of the infection with a sub-lethal innocua. In this model, it was unclear whether the loss of ADRB2 was responsible for hyperinflammation or whether it altered other aspects of the physiological response to infection. In general, administration of IL-10 to WT animals potently blocks the innate response leading to uncontrolled bacterial growth ^59,60, and blocking the IL-10R leads to hyper-inflammation and can protect animals from a lethal LM infection ⁶¹. We observed a significant increase of inflammatory cells into the spleens of Adrb2^−/− animals, which would suggest enhanced inflammation in the absence of ADRB2 signaling. However, we did not observe any significant differences in bacterial burden in either spleens or livers of Adrb2^−/− versus WT mice, which may indicate an ancillary role for ADRB2 signaling in this context.

We tested the involvement of the ADRB2 in the DSS-induced colitis model, which is very sensitive to regulation by IL-10 ^62,63. IL-10R on gut lamina propria macrophages is important for maintaining homeostasis, and its deletion leads to severe spontaneous colitis ⁶⁴ Here, we found that Adrb2^−/− animals lost weight more rapidly than their WT counterparts with a moderate reduction in overall colon length (although not statistically significant). This observation is in line with a role for IL-10 and IL-1 OR signaling in macrophages to modulate the severity of inflammation in this model⁶⁵

Finally, we found that the ADRB2 plays a critical role in LPS-induced sepsis. Adrb2^−/− mice were profoundly sensitive to an otherwise sublethal dose of LPS. This exquisite sensitivity correlated with a decrease in early systemic levels of IL-10 and increased TNF-α. The importance of IL-10 in ADRB2-mediated immune suppression was highlighted by 1) the rescue of Adrb2^−/− mice with a single dose of IL-10, 2) enhanced IL-10 expression in response to a β2-agonist, and 3) an increase in survival in response to a β2-agonist that was reversed by blocking the IL-1 OR. Finally, deletion of the Adrb2^−/− selectively in innate cells phenocopied the susceptibility to sepsis seen in the whole-body Adrb2-deficient strain, demonstrating a selective role for the ADRB2 in innate immune suppression.

Although adrenergic receptors have been shown to suppress LPS-driven TNF-a secretion and generally suppress inflammation, until now, there has been no comprehensive study probing the extent to which adrenergic signaling suppresses inflammatory processes emanating from such a broad spectrum of TLR agonists. Recent studies have attributed this to indirect mechanism in which NE promotes local secretion of Ach by T cells, which, in turn, can inhibit TNF-α secretion from macrophages ^66,67,68 While Ach has clearly defined immunosuppressive activities, our observation that selective deletion of the Adrb2^−/− on innate cells promoted lethal endotoxemia suggests an intrinsic role for E and NE in suppressing inflammation, which is distinct from the indirect actions of Ach.

Our findings have uncovered a unique synergistic pathway that converts an acute inflammatory signal into an anti-inflammatory response and likely explains a variety of phenomenon known to be involved in ADRB2-mediated immune suppression. For example, in cases of severe sepsis and septic shock in humans, NE is often administered in emergency situations as a vasopressor to restore blood pressure, heart rate, and lung function ^69,70 In addition to these physiological effects, NE administration has been shown to suppress TNF-α and increase serum IL-10 concentrations in a controlled LPS-induced model of sepsis in humans ⁷¹. There is, however, a trade-off to suppressing inflammation in this scenario, and that is the refractory condition known as immunoparalysis ⁷². This condition is characterized by prolonged immunosuppression where subjects fail to clear the original infection and present with increased vulnerability to subsequent secondary infections. Stimulation of β-adrenergic receptors is considered to be the causal factor in immunoparalysis and underscores the critical role played by this class of neurotransmitters in regulating immune function. As the ADRB2 is widely expressed on both innate and adaptive immune cells, the selective use of β-agonists and β-blockers is now on the forefront of therapies to both magnify responses to infections and tumors and to dampen such inflammatory processes in autoimmunity ¹⁵ Our finding that ADRB2 signaling directly controls IL-10 expression will lead to new avenues of immune control in a variety of therapeutic settings.

Highlights.

1. ADRB2 signaling synergizes with Toll receptors to drive rapid IL-10 secretion

2. ADRB2 deficiency increases inflammation to bacterial infection and by colitis

3. Loss of ADRB2 signaling in macrophages promotes lethal endotoxemia

4. Beta-agonists increase IL-10 and rescue animals from lethal endotoxemia