Chronic stimulation desensitizes β2‐adrenergic receptor responses in natural killer cells

Abstract

Adrenergic receptors (ARs) are preferentially expressed by innate lymphocytes such as natural killer (NK) cells. Here, we study the effect of epinephrine‐mediated stimulation of the β2‐adrenergic receptor (β2AR) on the function of human NK cells. Epinephrine stimulation inhibited early NK cell signaling events and blocked the function of the integrin LFA‐1. This reduced the adhesion of NK cells to ICAM‐1, explaining how NK cells are mobilized into the peripheral blood upon epinephrine release during acute stress or exercise. Additionally, epinephrine stimulation transiently reduced NK cell degranulation, serial killing, and cytokine production and affected metabolic changes upon NK cell activation via the cAMP‐protein kinase A (PKA) pathway. Repeated exposure to β2AR agonists resulted in the desensitization of the β2AR via a PKA feedback loop‐initiated G‐protein switch. Therefore, acute epinephrine stimulation of chronically β2AR stimulated NK cells no longer resulted in inhibited signaling and reduced LFA‐1 activity. Sustained stimulation by long‐acting β2‐agonists (LABA) not only inhibited NK cell functions but also resulted in desensitization of the β2AR. However, peripheral NK cells from LABA‐treated asthma patients still reacted unchanged to epinephrine stimulation, demonstrating that local LABA administration does not result in detectable systemic effects on NK cells.

β2‐adrenergic receptor (β2AR) stimulation of NK cells inhibits inside‐out signaling and LFA‐1 activation, explaining the mobilization of NK cells upon acute stress. Repeated or sustained β2AR stimulation desensitizes β2AR via G‐protein switch. Local administration of long‐acting β2‐agonists does not interfere with systemic NK cell functions in asthma patients.

Introduction

Natural killer (NK) cells are critical for early and effective immune responses against infections and cancer [1]. As part of the innate immune system, NK cells can kill transformed or infected cells and secrete cytokines such as IFNγ to help initiate and regulate adaptive immune responses [2]. Their activity is regulated by cytokines and a balanced array of activating and inhibitory surface receptors [3, 4]. Engagement of activating receptors such as NKp30, CD16, NKG2D, or 2B4 enables NK cells to kill targets by the release of lytic granules or by inducing death receptor‐mediated apoptosis [5].

NK cells express adrenergic receptors (ARs) and react to catecholamines. This is of physiological relevance for the neuro‐immune interaction for example during stress response [6, 7, 8]. Stress responses are mediated through two major axes: the fast sympathetic‐adrenomedullary and the delayed hypothalamus–pituitary–adrenal axis [9]. The sympathetic‐adrenomedullary axis activates the rapid sympathetic nervous system which results in the release of catecholamines, especially epinephrine. Known as the “fight or flight” response, this release induces a physical adaption to the threat such as pupil dilation, muscle contraction, elevated heart rate as well as activation and mobilization of immune cells [10]. Although these acute stress responses can boost the function of the immune system, chronic stress can result in immunosuppression [11, 12, 13] thus leading to higher incidence and progression of cancer [14, 15, 16].

Several in vitro studies demonstrated that epinephrine can inhibit NK cell functions [17, 18, 19]. This is in line with reports of decreased NK cell activity after stress from surgery and traumatic or thermal injuries [20, 21]. Therefore, blocking ARs during cancer surgery is suggested to have a beneficial effect for tumor control [22]. However, several in vivo studies analyzing psychological stressors or exercise demonstrate that the release of epinephrine leads to a redistribution of NK cells and an increase of NK cells in peripheral blood coinciding with elevated NK cell activities [23, 24, 25, 26].

The effects of epinephrine are mediated by the β2‐adrenergic receptor (β2AR), a G protein‐coupled receptor (GPCR) [27, 28]. The binding of β2AR agonists activates the Gαs subunits and leads to an increase of the intracellular cAMP concentration. The second messenger cAMP binds to protein kinase A (PKA) and activates the cAMP signaling pathway [28]. To turn off the β2AR response or desensitize the receptor due to a persistent stimulus, GPCR kinases can phosphorylate the receptor and initiate G‐protein independent signaling by β‐arrestin [29]. The binding of β‐arrestin to the β2AR can result in receptor internalization [28]. A second mechanism of GPCR desensitization is a switch in the G coupling subunits. A PKA feedback loop phosphorylates the β2AR and initiates a conformational change, reducing its affinity of the Gαs subunits and simultaneously increasing its binding affinity for the Gαi subunits, leading to an inhibited cAMP production [30].

Synthetic β2AR agonists are used as therapeutic drugs to treat asthma and chronic obstructive pulmonary disease (COPD) [31, 32]. Patients use long‐acting β2‐agonist (LABA) and short‐acting β2‐agonist (SABA) inhaler to relax the airway smooth muscle cells. However, regular LABA treatment is associated with tolerance and desensitization of β2AR which can decrease drug effectiveness [33].

Based on this knowledge, and due to the possible clinical relevance, our goal was to investigate the differences between acute and chronic effects of beta‐adrenergic agonists on NK cell functions. We also compared the effects of the natural agonist epinephrine as well as long‐acting compounds to better reproduce chronic stress and chronic beta‐adrenergic treatment, respectively.

Results

Epinephrine inhibits natural killer cell functions

Lymphocytes express alpha1 and beta subclass 2 AR [34], and the immune‐cell‐atlas of the ImmGen database shows that CD56⁺CD16⁺ NK cells express preferentially the β2AR. To confirm this, we analyzed the protein levels of the different ARs in PBMCs via flow cytometry. In comparison to B cells, T cells, and monocytes NK cells expressed high levels of the β2AR (Fig. 1A) but showed only low expression levels of the alpha1 subclasses 1A, 1B, and 1D (Fig. S1A). Furthermore, cultured human NK cells that were expanded in the presence of feeder cells and cytokines did not alter their β2AR expression (Fig. 1B). This demonstrates that ARs are preferentially expressed by innate lymphocytes such as NK cells. However, we did not test the expression of ARs on ILCs. Next, we wanted to test the functionality of these receptors by exposing freshly isolated or cultured NK cells to epinephrine. This inhibited the IFNγ secretion induced by CD16, NKp30, or NKG2D+2B4 engagement independently of NK cell cultivation (Fig. 1C). IFNγ secretion induced by other activation pathways such as K562 target cells or IL‐12/IL‐18 cytokine stimulation was also significantly inhibited by epinephrine (Fig. S1B). Similarly, epinephrine treatment inhibited NK cell degranulation as evident by reduced CD107a surface exposure upon engagement of CD16, NKp30, or NKG2D(+2B4) (Fig. 1D). This effect was stronger on freshly isolated NK cells. The addition of the beta blocker propranolol could reverse this inhibitory effect, which demonstrates that βARs are responsible for the effect of epinephrine on NK cells.

Figure 1.

β2‐Adrenergic receptor stimulation inhibits NK cell activation. (A) β2‐Adrenergic receptor expression analysis of PBMCs by flow cytometry (n = 8). Subsets were assigned according to the following markers: B cells (CD19⁺), NK cells (CD56⁺, CD3⁻), CD56 dim (CD56dim, CD3⁻), CD56 bright (CD56bright, CD3⁻), NK‐T cells (CD56⁺, CD3⁺), T cells (CD3⁺), and monocytes (FSC/SSC). (B) Representative β2‐adrenergic receptor (β2AR) expression histograms of freshly isolated or cultured NK cells. (C) IFNγ secretion of fresh NK cells (top) or cultured NK cells (bottom). NK cells were pretreated with epinephrine ± propranolol (each 1 µM) and stimulated for 5 h by plate‐bound antibodies as indicated. Supernatant was analyzed by IFNγ ELISA (mean, n = 3). (D) Degranulation of fresh NK cells (top) or cultured NK cells (bottom) was analyzed by CD107a expression. NK cells were pretreated and stimulated (3 h) like in (C), (mean, n = 3). Statistical analysis in (C) and (D) was performed using two‐way ANOVA test, ****p < 0.0001; ***p < 0.001; **p < 0.01; control set to 100%).

Physiologic plasma concentrations of epinephrine have been described to be in the low nanomolar range for nonstressed conditions. However, they can increase by a factor of 300 in stressful situations [35, 36]. Therefore, we titrated epinephrine and identified an inhibitory effect starting at 10 nM (Fig. S1C,D). However, as 1 µM resulted in more consistent results, we used this concentration for all subsequent experiments. Our results confirm the inhibitory effect of β2AR stimulation on NK cells in vitro [18, 19, 24] and demonstrate that its effect is independent of the culture condition or activation of NK cells.

β2‐Agonists block LFA‐1 activity and cytotoxicity of NK cells

In vivo, epinephrine release during acute stress, exercise, or medication can quickly and transiently increase NK cell numbers in peripheral blood [36, 37, 38]. This may be linked to changes in NK cell adhesion to endothelial cells [39] which is mediated by the integrin LFA‐1. The binding of LFA‐1 to its ligand ICAM‐1 can be stimulated by inside‐out signaling, which can be induced upon NK cell co‐activation via NKG2D+2B4 [40]. We analyzed this by ligand complex adhesion assay (LC‐AA) which determines the binding activity of LFA‐1 to its ligand ICAM‐1 [40, 41]. We found that epinephrine almost completely inhibits LFA‐1 activity, which could be reversed by the βAR antagonist propranolol (Fig. 2A). To further confirm the specific β2AR effect and to compare the natural agonist to a medically used agonist, we used the β2 sympathomimetic indacaterol. Indacaterol is a LABA, and it is proposed to dissociate slower from the membrane, to stay in close proximity to the β2AR, and to continuously stimulate the receptor [42, 43]. Like epinephrine, treatment with indacaterol significantly decreased LFA‐1 activity, which was also reversible by propranolol (Fig. 2A).

Figure 2.

β2‐Agonists block LFA‐1 activity and cytotoxicity of NK cells. (A) The LFA‐1 activity of cultured NK cells was measured by ligand complex adhesion assay (LC‐AA). NK cells were treated with epinephrine ± propranolol (left) or indacaterol ± propranolol (right) and stimulated by NKG2D/2B4 crosslinking. Medium served as control and DMSO as solvent control to indacaterol (n = 5). (B) The detachment of cultured NK cells from plate‐bound rhICAM‐1 was measured by xCELLigence RTCA. The cell index was normalized to the first addition time point. A representative curve is shown on the left. The normalized cell index of five independent experiments was quantified 40 min after β2‐adrenergic receptor (β2AR) agonist addition (right panel). (C) The cytotoxicity of NK cells + indacaterol or DMSO was evaluated by Incucyte microscopy against A549‐H2Bj‐eGFP cells. The specific lysis was determined after 8 h co‐incubation in an E:T ratio of 1.5:1. Results were normalized to A549 control (mean, n = 5). (D) Serial killing of cultured NK cells treated with indacaterol or DMSO was analyzed by live cell fluorescence microscopy. K562 and NK cells were co‐incubated for 16 h. A total of 100 individual NK cells out of four independent experiments were followed up and analyzed for noncytotoxic and cytotoxic contacts. Based on these data, serial killers (NK cells with more than one kill) were determined as a percentage of all NK cells that made at least one target cell contact (right panel). Statistical analysis was performed using ordinary one‐way ANOVA test or paired t‐test (in D) (****p < 0.0001; ***p < 0.001; **p < 0.01; *p < 0.05).

In the next step, we used a real‐time cell analyzer to measure the binding of NK cells to ICAM‐1. Adhesion of NK cells to ICAM‐1 coated plates was determined by changes in impedance which were analyzed over time in an xCELLigence device. Upon exposure to a β2AR agonist, we immediately detected NK cell detachment which peaked after 40 min (Fig. 2B). This LFA‐1 inhibition had a very fast onset and was similar for indacaterol and epinephrine. As NK cell adhesion and degranulation are essential events for NK cell cytotoxicity, we next determined the killing of tumor cells. Using a fluorescence microscopy‐based killing assay, we quantified the NK cell–mediated killing of the A549 lung cancer cell line. Treatment with the β2AR agonist led to a clear reduction of cancer cell killing (Fig. 2C). To additionally analyze if serial killing of NK cells was affected, we performed video microscopy and followed single NK cells after incubation with K562 target cells for 16 h. We then determined cytotoxic and noncytotoxic contacts for up to six target cell contacts per single NK cell as previously demonstrated [44, 45]. NK cells treated with indacaterol or solvent control had similar numbers of cell contacts. While we saw already a slight reduction in the first killing event for indacaterol‐treated NK cells, the number of NK cells showing serial killing activity was significantly decreased by indacaterol (Fig. 2D). These data demonstrate that stimulation of β2AR in vitro inhibits NK cell cytotoxicity and particularly affects the serial killing activity of these cells. Additionally, the loss of LFA‐1 activity may explain the increased numbers of NK cells in the blood stream upon epinephrine release.

β2AR stimulation acts via cAMP‐PKA pathway and blocks the phosphorylation of activating receptors

Next, we wanted to investigate the signaling of the β2AR in NK cells. Stimulation of the G protein‐coupled β2AR results in the activation of adenylate cyclase (AC) [46] followed by cAMP production. cAMP can then activate PKA, which can phosphorylate transcription factors or cytoplasmatic enzymes [47]. Additionally, cAMP can also regulate the function of ion channels or bind to the exchange protein directly activated by cAMP (EPAC) [48]. We found significantly increased levels of cAMP in NK cells after epinephrine stimulation (Fig. 3A). Using the LC‐AA as functional readout, we analyzed the role of the cAMP downstream molecules PKA and EPAC. NK cells were treated with ESI09 (EPAC inhibitor), H89 (PKA inhibitor), or with a combination of both inhibitors. While ESI09 only slightly affected the inhibition of LFA‐1 binding activity induced by epinephrine, H89 could significantly reverse the inhibitory effect of epinephrine. The combination of both inhibitors showed almost no further increase in LFA‐1 activity, demonstrating that the inhibitory effect of epinephrine was mostly mediated by PKA (Fig. 3B). To understand where β2AR stimulation interrupted the activating signals of NK cells, we analyzed the phosphorylation of the activating receptor 2B4 and the downstream proteins Vav and ERK. The Western blot analyses showed that epinephrine (Fig. 3C–E) and indacaterol (Fig. S2) interfered with the stimulation‐dependent phosphorylation of 2B4 and subsequently led to reduced phosphorylation of Vav and ERK. This inhibition could be reversed by propranolol. This demonstrates that the β2AR acts mainly via the cAMP‐PKA pathway and interferes with NK cell signaling already at the level of activating receptor phosphorylation.

Figure 3.

β2‐Adrenergic receptor (β2AR) stimulation acts via cAMP‐protein kinase A (PKA) pathway and blocks the phosphorylation of activating receptors and signaling molecules. (A) The cAMP concentration in natural killer (NK) cells was evaluated by ELISA. NK cells were pretreated with epinephrine (1 µM) or forskolin (100 µM) as positive control before lysis. (B) The LFA‐1 activity was measured by LC‐AA. NK cells were preincubated with PKA inhibitor (H89, 10 µM) and/or EPAC inhibitor (ESI09, 25 µM). NK cells were treated with epinephrine (1 µM) ± inhibitor (were kept throughout the whole experiment). In (C), the phosphorylation of the 2B4 receptor was measured via immune precipitation and Western blot. NK cells were activated via cross‐linking NKG2D/2B4 and treated with epinephrine ± propranolol (1 µM) for 5 min. In (D) and (E), the phosphorylation of downstream molecules Vav and ERK was measured by Western blot. The densitometry was quantified via ImageJ. Statistical analysis was performed using paired t‐test for cAMP ELISA (A) or ordinary one‐way ANOVA test (****p < 0.0001; ***p < 0.001; **p < 0.01; *p < 0.05). Controls were set to 1.

β2AR stimulation changes the metabolic profile of activated NK cells

As acute β2AR stimulation led to an inhibition of NK cells, we also wanted to investigate if β2AR influences the metabolic profile of activated NK cells. Therefore, we utilized the Seahorse analyzer and its extracellular flux rate technology to measure the oxygen consumption rate (OCR) and the extracellular acidification rate (ECAR), which correlate with mitochondrial respiration and glycolysis, respectively [49]. Stimulation via anti‐CD16 antibody increased OCR and ECAR levels of NK cells (Fig. 4), consistent with increased metabolism due to NK cell activation. Treatment with epinephrine or indacaterol before CD16 stimulation could inhibit the increase in OCR levels, and this effect was reverted by propranolol (Fig. 4A,B top panels).

Figure 4.

β2‐Adrenergic receptor (β2AR) stimulation changes the metabolic profile of CD16 activated NK cells. The metabolic profile of cultured NK cells was evaluated by Seahorse technology. Representative measures are shown for oxygen consumption rate (OCR, top) and extracellular acidification rate (ECAR, bottom). Each condition was analyzed in triplicates. NK cells were treated in three injection steps. First step: propranolol or medium, second step: medium, epinephrine (A) or DMSO, indacaterol (B), third step: activation via CD16 antibody addition. The oxidative phosphorylation (OCR) and glycolysis (ECAR) measurements were normalized to the point of β2AR agonist addition and quantified 12 min after stimulation (TStim) or at the end (TEnd) (right panels, n = 6 in A, n = 4 in B). Statistical analysis was performed using ordinary one‐way ANOVA test (****p < 0.0001; ***p < 0.001; **p < 0.01; *p < 0.05).

The injection of epinephrine or indacaterol induced a distinct but transient increase in ECAR levels (Fig. 4A,B). Although the acute increase in ECAR levels after CD16 stimulation was only slightly reduced by the β2AR agonists, they clearly prolonged the increase in ECAR levels (Fig. 4), suggesting that β2AR stimulation increases glycolysis in NK cells. This was also observed when we stimulated NK cells via IL‐12/IL‐15/IL‐18 (Fig. S3). PKA is described to modulate glycolysis and increase glucose uptake by glucose transporters [50]. However, the PKA inhibitor H89 by itself already increased ECAR levels in NK cells, making it difficult to determine if PKA inhibition could abolish the β2AR‐induced prolongation of increased ECAR values (Fig. S4). Additionally, we could not observe an upregulation of glucose transporters in NK cells after epinephrine stimulation (Fig. S5). This demonstrates that the engagement of β2ARs inhibits the increase in mitochondrial respiration upon CD16 stimulation, whereas it prolonged the CD16‐induced increase in glycolysis, by a so far unknown mechanism.

Chronic stimulation desensitizes β2AR signaling and agonist induced inhibition

Chronic stress leads to repeated increases in epinephrine concentrations and negative effects on cancer patients and on the recovery after surgery [51, 52]. Therefore, we wanted to understand if repeated administration of β2AR agonists also leads to the same inhibitory effects on NK cells as observed during the acute situation. We stimulated NK cells for 5 consecutive days with epinephrine or indacaterol and analyzed NK cell functions after the last stimulation (Fig. 5A). In contrast to acute stimulation, the repeated exposure to β2AR agonists showed no inhibitory effect on NK cells (Fig. 5). CD16‐induced IFNγ secretion and NK cell degranulation as well as LFA‐1 activity were no longer inhibited by epinephrine or indacaterol when NK cells had been chronically stimulated via β2AR for 96 h (Fig. 5B–D). This was also observed when we investigated NKp30 or NKG2D+2B4‐induced NK cell degranulation (Fig. S6A). Similarly, real‐time detachment analysis by xCELLigence demonstrated that NK cells only detach from the ICAM‐1 coated plates upon acute β2AR stimulation but do no longer react after chronic stimulation (Fig. 5E). Killing of A549 tumor cells by chronically stimulated NK cells was similar to control‐treated cells and was no longer inhibited by indacaterol (Fig. 5F). Acute β2AR stimulation of chronically epinephrine or indacaterol‐treated NK cells no longer showed any inhibition of ERK and Vav phosphorylation (Fig. 5G, Fig. S6B). Finally, the metabolic profile of chronically β2AR‐treated NK cells was unaltered when acutely stimulated by β2‐agonist (Fig. S7). This demonstrates that chronic β2AR stimulation desensitizes NK cells to the inhibitory effect of acute β2AR stimulation on NK cell signaling, metabolism, and function.

Figure 5.

Chronic stimulation desensitizes β2AR and abolishes NK cell inhibition. (A) Cultured NK cells were either treated chronically every 24 h for five consecutive times or acutely with 1 µM epinephrine/indacaterol. (B) The IFNγ secretion was measured by ELISA. NK cells were stimulated (6 h) with plate‐bound CD16 antibody (median, n = 4–5). (C) The degranulation was measured by flow cytometric CD107a expression analysis. NK cells were stimulated for 3 h like in (B) (median, n = 5). (D) LFA‐1 activity was analyzed by LC‐AA (n = 7). (E) The detachment of chronically or acutely treated NK cells from plate‐bound rhICAM‐1 was measured by xCELLigence RTCA. The cell index was normalized to the time point of addition. A representative curve is shown on the left. The right panel shows the quantification 40 min after normalization of three independent experiments (median, n = 3). (F) The cytotoxicity of chronic stimulated NK cells was evaluated by Incucyte microscopy against A549‐H2Bj‐eGFP cells. The specific lysis was determined after 8 h co‐incubation in an E:T ratio of 1.5:1. Results were normalized to A549 control (mean, n = 8). (G) Phosphorylation of ERK was analyzed by Western blot. NK cells were acutely or chronically treated with 1 µM indacaterol before analysis. Protein bands were quantified by ImageJ (median, n = 4, controls set to 1). Statistical analysis was performed using ordinary one‐way ANOVA (in F) or two‐way ANOVA test (****p < 0.0001; ***p < 0.001; **p < 0.01; *p < 0.05).

Chronic stimulation desensitizes the β2AR via PKA feedback loop and β2AR agonists show different stimulation kinetics

Next, we wanted to investigate why long‐term β2AR stimulation attenuates the inhibitory effect of acute β2AR stimulation on NK cells. We repeatedly treated NK cells with epinephrine or indacaterol for 96 h and performed RNAseq analysis using solvent‐treated NK cells as a control. Among the ∼16 500 identified genes only very few were differentially regulated, and we found two downregulated and no upregulated genes for indacaterol‐treated NK cells (Fig. 6A) as well as three downregulated and one upregulated gene for epinephrine‐treated NK cells (Fig. S8A). However, apart from one gene (PPFIA4), there was no overlap of the differentially regulated. This demonstrates that changes in gene transcription are unlikely the reason for the effect of long‐term β2AR stimulation on NK cells.

Figure 6.

Chronic β2AR stimulation does not alter transcription levels or protein expression. NK cells were chronically (96 h) treated with 1 µM indacaterol and analyzed by RNAseq (A) or proteomics (B). Representative volcano plots are shown (n = 3). (C) Representative histogram of the β2‐adrenergic receptor (β2AR) expression of untreated NK cells or 96 h treated with indacaterol. Expression was determined by flow cytometry. (D) The cAMP concentration was measured of chronically or acutely stimulated (indacaterol) NK cells ± Gαi inhibitor (Pertussis Toxin, PTX). The cAMP concentration was analyzed by ELISA of lysed NK cells (mean, n = 3). In (E), NK cells were treated with indacaterol (1 µM) and restimulated at indicated time points. The LFA‐1 activity was analyzed by LC‐AA. In (F), NK cells were treated for 24 h with DMSO (left panel) or indacaterol (right panel) in the presence of the protein kinase A (PKA) inhibitor H89, the β‐arrestin/β2‐adaptin interaction inhibitor barbadin or the respective solvent controls (medium for H89 and DMSO for barbadin). After 24 h, the cells were washed and restimulated with DMSO (left side of the graphs) or indacaterol (right side of the graphs), and LFA‐1 activity was determined by LC‐AA assay. Medium control was set to 1 and compared to samples (median, n = 6). (G) NK cells were pretreated with DMSO (pre‐DMSO, blue), Epi (pre‐Epi, red), or Inda (pre‐Inda, yellow) and then washed and incubated in media for 3 days. Subsequently, the cells were treated a second time with medium, epinephrine, or indacaterol, and the LFA‐1 activity of NK cells was analyzed by LC‐AA (median, n = 8). (H) NK cells were pretreated for 2 min with medium or epinephrine (Epi) and immediately analyzed by LC‐AA (left panel, median, n = 6) or CD16‐induced degranulation (CD107a, right panel, median, n = 3). Epi‐treated cells were also washed and then again treated with medium (Epi + wash‐medium) or epinephrine (Epi + wash‐Epi) and analyzed. Statistical analysis was performed using ordinary one‐way ANOVA (in E–G) or two‐way ANOVA test (in D) (****p < 0.0001; ***p < 0.001; **p < 0.01; *p < 0.05).

To test for protein expression, we performed proteomic analysis for long‐term β2AR stimulated NK cells (Fig. 6B, Fig. S8B). Among the >1100 identified proteins, we only found three proteins to be differentially regulated in the indacaterol‐treated NK cells (Fig. 6B) and only seven differentially regulated proteins in the epinephrine‐treated NK cells (Fig. S8B), again with no overlap. Therefore, changes in protein expression levels are also unlikely to be the cause for the effect of long‐term β2AR stimulation. Consistent with this, we also found no changes in β2AR expression levels in NK cells upon chronic β2AR agonist stimulation (Fig. 6C, Fig. S8C).

However, cAMP concentrations were no longer increased when comparing NK cells after chronic versus acute indacaterol treatment (Fig. 6D). Interestingly, cAMP levels upon chronic stimulation showed a tendency to be even lower than in control‐treated NK cells, suggesting inhibitory signaling by chronically stimulated β2ARs. In line with this, we found that the significant difference between acute and chronic treatment was abolished when NK cells were additionally treated with the Gαi protein inhibitor pertussis toxin (PTX) (Fig. 6D). This suggests that chronic β2AR stimulation induces a switch from activating Gαs to the inhibitory Gαi.

By analyzing LFA‐1 activity at various time points after indacaterol addition, we could identify that NK cells start to become desensitized already 4 h after indacaterol exposure. NK cells were fully desensitized after 24 h of indacaterol treatment (Fig. 6E). PKA‐mediated phosphorylation of β2AR was described as a feedback loop that initiates a switch of coupling from activating Gαs to the inhibitory Gαi [53]. To analyze this, we tried to interfere with the indacaterol‐induced desensitization by blocking PKA via the inhibitor H89. We treated NK cells with DMSO as solvent control or indacaterol in the presence or absence of H89 for 24 h. Cells were then washed and stimulated again with DMSO or indacaterol, and LFA‐1 activity was determined by LC‐AA assay. In DMSO‐pretreated NK cells, the acute stimulation with indacaterol showed the expected inhibitory effect (Fig. 6F, left panel). This was no longer observed when NK cells had been pretreated with indacaterol, consistent with a desensitization of the β2AR (Fig. 6F, right panel). Interestingly, H89 treatment during the indacaterol pretreatment reestablished the inhibitory effect of acute indacaterol exposure (Fig. 6F, right panel) suggesting that PKA activity is necessary for the desensitization of the β2AR. In the same assay, we analyzed the effect of Barbadin, a selective β‐arrestin/β2‐adaptin interaction inhibitor. Barbadin and its DMSO control already showed a significant effect on the LFA‐1 activity in the control group, suggesting that high amounts of DMSO impacted the LFA‐1 activity (Fig. 6F, left panel). However, Barbadin showed no significant effect on the desensitization when co‐incubated with indacaterol (Fig. 6F, right panel), suggesting that β‐arrestin is not involved in the desensitization of the β2AR.

To identify how long the desensitizing effect lasts, we treated NK cells with DMSO, epinephrine, or indacaterol and subsequently washed the agonists away. After 3 days, the LFA‐1 activity was determined after restimulation with epinephrine or indacaterol. Although the LFA‐1 activity of NK cells pretreated with epinephrine was inhibited when restimulated by a β2‐agonist (Fig. 6G, red symbols), NK cells pretreated with indacaterol were still desensitized (Fig. 6G, yellow symbols). To test how long the acute effect of β2AR stimulation lasts, we treated NK cells with epinephrine, washed the cells, and then again treated them with medium or epinephrine. Although acute epinephrine treatment inhibited LFA‐1 activity and degranulation, this was no longer observed after washing the cells, demonstrating that this inhibition is very transient (Fig. 6H). We also did not observe any desensitization of the β2AR as epinephrine treatment of the washed cells still resulted in inhibition of LFA‐1 activity and degranulation.

Peripheral NK cells from asthma patients persisted to react to epinephrine

β2‐Agonists are also used as therapeutics with a daily application [33]. In respiratory diseases such as asthma or COPD, β2AR agonists and specifically LABA inhalers are the mainstream treatment for these patients [54]. Therefore, we investigated if long‐acting β2AR agonist treatment may have systemic effects on NK cells in vivo. We isolated PBMCs from 10 asthma patients and 10 healthy control subjects (Fig. 7A). Patients had moderate to severe asthma and used salmeterol (3) or formoterol (7) for several years. Like indacaterol, salmeterol and formoterol are LABAs with lipophilic structures as amino substituent and the duration of action lasts for more than 12 h [55, 56]. The NK cell frequency of patients and healthy controls were similar (Fig. S9A) and showed no significant differences in the expression of activating receptors (Fig. S9B). To analyze if β2ARs were desensitized by the long‐acting β2AR agonists, we determined the CD16‐induced IFNγ secretion, NK cell degranulation, and LFA‐1 activity of asthmatic and healthy control NK cells (Fig. 7B–D). For all these parameters, NK cells from healthy subjects and asmathics showed similar inhibitory reactions to epinephrine treatment. This suggests that chronic inhalation of LABA drugs does not result in a desensitization of β2ARs on peripheral blood NK cells.

Figure 7.

Peripheral NK cells from asthma patients and healthy controls show comparable reactions to epinephrine. (A) Tables of cohort data (top) and clinical characteristics of asthma patients (bottom). PBMCs of participants were isolated and further analyzed by functional assays. (B) PBMCs were treated with epinephrine and stimulated by plate bound CD16 antibody (6 h). IFNγ secretion was measured by ELISA (n = 10). (C) Like in (B), PBMCs were treated with epinephrine and stimulated via CD16 (3 h) to measure degranulation frequency (CD107a⁺) of NK cells by flow cytometry (n = 10). (D) The LFA‐1 activity of NK cells was determined by LC‐AA. PBMCs were treated with epinephrine (1 µM) and stimulated by crosslinking NKG2D/2B4 antibodies (n = 9). In the left panels, the data points from the same donor are connected. Right panels, statistical analyses using a two‐way ANOVA test (**p < 0.01, *p < 0.05).

Discussion

Among all lymphocytes, NK cells show the highest expression of the β2AR, explaining why these innate lymphocytes respond strongly to stress and sympathetic nervous system activation. It is known that acute stress or physical exercise mobilizes leukocytes in the bloodstream [57], which can be detected within minutes. The release of catecholamines flushes preferentially protective, cytotoxic immune cells out of the marginal pool into the circulation [10, 37–39]. Our data show that epinephrine can very quickly reduce the binding activity of the integrin LFA‐1 on NK cells, which may explain their fast detachment from endothelial cells and explain their mobilization during acute stress. Mechanistically, LFA‐1 activity is dependent on inside‐out signaling by activating NK cell receptors [40]. The inhibition of early signaling events such as the phosphorylation of activating receptors by epinephrine may explain how β2AR stimulation can block inside‐out signaling and thereby reduce LFA‐1 binding activity.

Our data show that acute epinephrine exposure inhibits IFNγ secretion, degranulation, and serial killing activity of NK cells. Mechanistically, this was dependent on β2AR‐mediated cAMP production and PKA activation, which interferes with early signaling events in NK cells such as 2B4, ERK, and Vav phosphorylation. The acute effect of epinephrine on NK cell cytotoxicity has been analyzed before. Interestingly, although some studies showed an inhibitory effect of epinephrine on NK cells [18, 19, 58], other in vivo studies demonstrated an increase in NK cell cytotoxicity [24, 26, 37]. In line with an increased cytotoxicity, our metabolic analysis discovered a prolonged CD16‐induced glycolytic activity after β2AR stimulation. Segerstrom and Miller concluded in a meta‐analysis that most increased activity measurements in vivo were based on increased NK cell numbers in the periphery rather than on an increased cytotoxic activity of individual NK cells [13]. A further explanation for the discrepancy of results is the timing of measurements. Our data show that the inhibitory effect of epinephrine only lasts for a few minutes. This transient inhibition may be sufficient to reduce LFA‐1 activity, break NK‐endothelial cell interactions, and mobilize NK cells in the bloodstream. However, once mobilized the inhibition of NK cell activities would quickly wear off resulting in increased NK cell numbers with normal functions. Therefore, depending on the time lapse between stress exposure and NK cell isolation in the different studies, the inhibitory effect of epinephrine could be still active or already lost, thus leading to the contradictory results. Importantly, the short inhibitory effect of epinephrine in the circulation shows that the main task of the hormone is not to inhibit effector functions, but rather to recruit cytotoxic cells as immunological defense. This effect holds great potential. For example, exercise‐mediated tumor immunotherapy is considered part of a cancer therapy strategy [59]. Besides the general benefit of physical activity, Pederson and colleagues demonstrated that running suppressed tumor growth through epinephrine and IL‐6‐dependent NK cell mobilization in mouse models [60]. A similar observation was made in humans where exercise training was reported to increase NK cell cytotoxicity in breast cancer and stomach cancer patients [61, 62, 63].

In contrast to the effects of acute stress, chronic stress has negative effects on the immune system and can increase the risk of diseases, infections, and cancer [11, 64–66]. However, the mechanism of how chronic stress increases disease susceptibility remains unclear. Most studies on NK cells show a reduced activation and cytotoxicity upon chronic stress. In animal models, chronic epinephrine led to progression of leukemia or to an increased risk to MCMV infection [7, 67]. Our data demonstrate that sustained β2AR stimulation by using the LABA indacaterol results in decreased NK cell cytotoxicity and serial killing. Therefore, β2AR stimulation during chronic stress would inhibit NK cell functions. Indeed, acute stress in chronically stressed people induced the release of higher epinephrine levels and resulted in a more pronounced reduction of NK cytotoxicity [68].

Upon repeated β2AR stimulation of NK cells, we found that acute epinephrine treatment no longer resulted in reduced LFA‐1 activity and inhibited NK cell effector functions. Mechanistically, this was mediated by a PKA feedback loop initiating a G‐protein switch at the receptor. PKA‐induced phosphorylation of β2AR changes the conformation of the receptor resulting in a higher affinity for Gαi subunit binding [30]. Therefore, the stimulation of the β2AR in chronic epinephrine or indacaterol‐treated NK cells resulted in a Gαi‐mediated inhibition of the AC, explaining the reduced levels of cAMP. We confirmed the G‐protein switch by PTX‐induced blockade of the Gαi‐coupled receptors resulting in increased cAMP levels. In vivo, such a desensitization would result in a lack of NK cell mobilization, as LFA‐1 activity is increased by Gαi‐coupled GPCRs [69], which may stabilize NK‐endothelial cell interactions. This reduced NK cell mobilization could explain the reduced NK cell activity observed upon chronic stress in vivo.

Our data show that sustained stimulation by LABA leads to a desensitization of the β2AR within 24 h. Similar effects have been shown for LABA therapies resulting in the loss of drug effectiveness [33, 70]. LABA‐mediated desensitization of β2AR on airway smooth muscle cells can induce tolerance to SABA treatments which led to decreased asthma control and hospitalization [71]. In line with our results, such desensitization and tolerance to β2‐agonists can be reached within the first days of therapy [72]. Our results suggest that LABA treatment may result in desensitized or inhibited NK cells. The analysis of peripheral NK cells from asthma patients showed no difference to healthy controls in response to epinephrine stimulation, demonstrating that local LABA administration does not result in detectable systemic effects. However, we were unable to investigate if NK cells in the lung of these patients were affected by the long‐term LABA treatment. As the airway is a main pathogenic entry site which is also controlled by NK cells [73], desensitization and a Gαi‐coupled β2AR could potentially dysregulate NK cell migration and recruitment, thereby impairing immune responses.

In summary, we show that acute β2AR stimulation leads to a reduced LFA‐1 activity and a transient inhibition of NK cells whereas chronic stimulation desensitizes the β2AR. Although already a single treatment with the LABA indacaterol induces desensitization in vitro, LABA therapy did not affect peripheral NK cells of asthma patients.

Materials and methods

Isolation of NK cells and cell culture

Human NK cells were isolated from PBMCs with Dynabeads Untouched Human NK Cell‐Kit according to the manufacturer's instructions (Thermo Fischer Scientific). For experiments with fresh NK cells, isolated NK cells were rested in IMDM GlutaMAX medium (Thermo Fischer Scientific), 10% FCS, 1% penicillin/streptomycin (Thermo Fischer Scientific) overnight and then used for experiments. NK cell isolation purity and viability (7AAD, BioLegend) were confirmed by flow cytometry. NK cells were defined as CD3⁻ and CD56⁺ cells. To culture cells, isolated NK cells were seeded in 96‐well round‐bottom plates (Nunc) at a density of 1.5 × 10⁶/mL with irradiated feeder cells (K562‐mbIL15‐mbIL21‐41BBL) and 200 U/mL IL‐2 (NIH Cytokine Repository). On Day 8, NK cells were restimulated with 200 U/mL IL‐2. In the next weeks, NK cells were split to a density of 1.5–2 × 10⁶/mL, cultured in the presence of 100 U/mL IL‐2, and after 4 weeks used as preactivated, cultured NK cells.

Cell lines

K562 cells were cultured in IMDM medium (Thermo Fischer Scientific), 10% FCS, 1% penicillin/streptomycin. A549‐H2Bj‐GFP cells have been kindly provided by Dr. Slava Ziegler and were cultured in DMEM (Thermo Fischer Scientific), 10% FCS, 1% penicillin/streptomycin.

Acute and chronic β2AR stimulation

NK cells were treated with epinephrine (Sigma‐Aldrich) or indacaterol (Cayman Chemicals) at a concentration of 1 µM. Acutely stimulated NK cells were treated for 0.5 h before or directly in the experiment. For chronic β2AR stimulation, NK cells were treated every 24 h for a total of 96 h with a β2AR agonist.

Flow cytometry

Protein expression levels were evaluated using the BD LSRFortessa flow cytometer according to published guidelines [74]. PBMC and NK cell AR expression levels were examined by incubation with alpha‐1A AR (PolyAb, Proteintech), alpha‐1B (471802, Invitrogen), alpha‐1D AR (PolyAb, Invitrogen), or β2AR (6H8, Abcam) at RT for 20 min followed by secondary antibody staining (goat‐anti‐mouse PE or goat‐anti‐rabbit PE [Jackson ImmunoResearch]). Cells were stained with the following reagents: 7AAD (Invitrogen), CD16‐PE (3G8, Biolegend), GLUT‐1 (202915, R&D systems), GLUT2‐AF647 (199017, R&D systems), and GLUT3‐AF700 (202017, R&D systems). Data were analyzed using FlowJo (version 10) software.

IFN‐γ ELISA, degranulation assay

Nunc MaxiSorp plates were prepared by antibody immobilization. Indicated antibodies (NKG2D [149810; R&D Systems], 2B4 [C1.7; Beckman Coulter], NKp30 [produced in our lab], CD16 [3G8, BioLegend]) were incubated overnight at 4°C at a concentration of 1 µg/mL. Next day, 0.2 × 10⁶ NK cells (fresh or cultured) were pretreated with epinephrine or indacaterol (1 µM) for 0.5 h and subsequently stimulated with plate‐bound antibodies or IL‐12/‐18 (0.5 ng/mL, R&D/2.5 ng/mL, MBL Life Science) for 5 h at 37°C, 5%CO₂. The supernatant was collected after centrifugation (5 min, 500 × g) and stored at −20°C. The secreted IFN‐γ concentration was determined by human ELISA MAX Deluxe Set Human IFN‐γ from BioLegend according to manufacturer's instructions. For degranulation assay, NK cells were pretreated with β2AR agonists and activated by plate‐bound antibodies. During 3 h incubation, NK cells were stained with anti‐CD107a‐PE‐Cy5 (H4A3, Biolegend) and analyzed by flow cytometry.

IncuCyte S3 microscopy‐based killing assay

IncuCyte S3 live‐cell analysis was conducted at 37°C and 5% CO₂ using 10× magnification. A549‐H2Bj‐GFP cells were seeded in a density of 3.5 × 10^3 cells/well and incubated in RPMI1640 media overnight. On the second day, pretreated NK cells were added in an E:T ratio of 1.5:1. NK cells were either acutely or chronically treated with indacaterol (1 µM) or DMSO (Thermo Fischer Scientific) as solvent control. After 8 h of co‐incubation, the plates were washed with PBS and quantitatively analyzed for GFP fluorescence by IncuCyte 2019B Rev Software. A549‐H2Bj‐GFP without NK cells were used as control and for calculation of the specific lysis:

$$\%\mathrm{Specific}\mathrm{lysis}=100\times \frac{\left(\mathrm{A}549+\mathrm{NK}\right)-\left(\mathrm{A}549\mathrm{control}\right)}{0-\left(\mathrm{A}549\mathrm{control}\right)}$$

Serial killing

Microscope‐based serial killing analysis has been performed as described [75]. Briefly, K562 cells were seeded on a microchip with SYTOX blue dead cell stain (1 µM, Thermo Fischer Scientific) in IMDM medium. NK cells were stained with CellTracker Red (5 µM, Thermo Fischer Scientific), pretreated with DMSO or indacaterol (1 µM), and subsequently added to the microchip. Time‐lapse live‐cell microscopy was immediately started using a Zeiss Axio Observer Z1 7 microscope equipped with a 20×/0.8 Plan‐Apochromat objective and an incubation chamber with environmental control (37°C, 5% CO₂, and humidity device S1). Images were acquired every 3 min for 16 h.

Ligand complex adhesion assay (LC‐AA)

The assay was performed essentially as described [40]. Briefly, ICAM‐1–Fc complexes were prepared by mixing 50 ng/mL recombinant human ICAM‐1–Fc chimera (R&D Systems) and F(ab)2 fragments of goat anti‐human Fcγ fragment specific antibody (80 ng/mL, PE labeled; Jackson ImmunoResearch) in buffer without cations for ≥20 min at room temperature. NK cells were incubated with 1 µg/mL anti‐NKG2D (149810; R&D Systems) and anti‐2B4 (C1.7; Beckman Coulter). After washing, cells were resuspended in buffer containing indacaterol (1 µM), DMSO or epinephrine (1 µM), and ICAM‐1–Fc complexes were added (dilution 1:20). A goat anti‐mouse IgG (Dianova) was added to a final concentration of 2.3 ng/mL for Ab cross‐linking. Cells were fixed and analyzed by flow cytometry. When indicated, the cells were additionally treated with inhibitors (10 µM H89 [Cayman Chemicals] or 25 µM ESI09 [TOCRIS]). In restimulation experiments, inhibitors (10 µM H89 or Barbadin 50 µM [MedChemExpress]) were kept in media for 24 h and were washed away before LC‐AA and second β2AR stimulation.

xCELLigence‐based detachment assay

E‐plates were coated with 7 ng/mL goat anti‐mouse antibody (Dianova) followed by 2 ng/mL of recombinant human ICAM‐1–Fc chimera (R&D Systems). After the addition of 100 000 NK cells per well, impedance was measured for 2 h in 10 min intervals. When the cell index reached a plateau, epinephrine, indacaterol (1 µM), or medium, DMSO was added, and the cell index was determined every 2 min. For the analysis, the cell index was normalized to the point of β2AR stimulation. Experiments were performed on xCELLigence RTCA DP.

cAMP ELISA

Cultured NK cells were chronically stimulated with indacaterol (1 µM) or DMSO for 96 h. When indicated, NK cells were pretreated with PTX (100 ng/mL [TOCRIS]) for 30 min at 37°C before β2AR stimulation. After 95 h, NK cells were incubated for 0.5 h in 100 µM IBMX (Cayman Chemicals) to inhibit phosphodiesterases before last indacaterol incubation step (0.5 h). 1 × 10⁶ NK cells were lysed in 100 µL 0.1 M HCL, and the supernatant was stored at −20°C until cAMP analysis. The ELISA was performed according to manufacturer Cayman Chemicals instructions.

Immunoprecipitation and Western blot

For 2B4 phosphorylation analysis, 4 × 10⁷ NK cells were incubated for 20 min at room temperature with 2B4/NKG2D antibodies (10 µg/mL). After a washing step, NK cells were treated with PBS or epinephrine ± propranolol (2 µM each) and activated by crosslinking the antibodies by a goat anti‐mouse antibody (10 µg/mL [Dianova]) for 0 or 5 min at 37°C. The reaction was stopped by the addition of ice‐cold PBS. Cells were washed and lysed, and the supernatant was transferred to pan mouse IgG beads (approx. 7 × 10⁶) and rotated for 1 h at 4°C. Beads were washed in lysis buffer and dried. Dried beads were shortly boiled for 5 min at 95°C in RSB buffer, and solution was used for SDS PAGE. Western blotting was essentially performed as described [76].

Metabolic extracellular flux assays

The Seahorse XFe96 Analyzer and Wave software were used to measure the NK cell glycolysis (ECAR) and oxidative phosphorylation (OCR) via extracellular flux. Cultured NK cells (250 000 cells/well) were immobilized for 30 min at 37°C without supplemented CO₂ in Seahorse RPMI media (10 mM glucose, 1 mM pyruvate, 2 mM glutamine, Agilent Technologies) on poly‐(l)‐lysine (Sigma‐Aldrich) coated cell culture plates. The measurement steps were set to a 3 min mixing/3 min measurement interval. Epinephrine (1 µM), indacaterol (1 µM), DMSO, or propranolol (1 µM) were injected as indicated before NK cells were activated by CD16 antibody (1 µg/mL, 3G8) or cytokines IL12/15/18 (50 ng/mL (R&D Systems)/250 ng/mL (Pan Biotech)/250 ng/mL MBL Life Science). After 160 min, the measurements were terminated and normalized to the time point of β2AR agonist addition.

Transcriptomics

Cultured NK cells from three different donors were chronically stimulated for 96 h with epinephrine (1 µM), indacaterol (1 µM), or DMSO as described above. NK cells treated with medium were used as a reference. The transcriptomic analysis was carried out by Eurofins GATC Biotech GmbH. The expression analyses of the RNA‐Seq reads were aligned to the reference transcriptome using Bowtie alignments. The variant analysis was done by GATK Haplotype Caller. The aligned reads were used by MATS to detect alternative splicing events. The R package GenomicFeatures [77] was used to summarize transcript reads at gene level. For prefiltering, genes with a mean count of less than 1 read across the nine samples (6893) were removed, leaving 16 566 genes for further analysis. Differential gene expression analysis was performed using the R package DESeq2 [78]. A general linear model, including the variables replicate and treatment, was fitted to determine differentially expressed genes (DEGs). DEGs were then calculated for the comparisons “epinephrine vs. untreated” and “indacaterol vs. untreated,” respectively. For more reliable effect estimates, adaptive shrinkage was applied [79]. This leads to shrinkage of log₂‐transformed fold‐changes (log₂FCs) toward zero if expression changes are mostly due to noise, whereas relevant log₂FCs are preserved. For each comparison, a gene was considered to be differentially expressed if the effect size satisfies log₂FC > 1 for upregulation (log₂FC < −1 for downregulation), and the estimate is significantly different from zero (i.e., no effect) with a false discovery rate (FDR)‐adjusted p value p_adj  < 0.05.

Proteomics‐LC‐MS/MS

Cultured NK cells (10⁶ cells per sample) were chronically stimulated for 96 h with epinephrine (1 µM) or indacaterol (1 µM). NK cells only treated with medium were used as a reference. After chronic stimulation, supernatant was discarded, and cells were frozen at −80°C in 10%DMSO/FCS. Cryocultures were resuspended in 80% methanol, lysed at 4°C by ultrasonication (Bioruptor, Diagenode), and centrifuged (21 000 × g, 5 min, 4°C). Next, pellets were proteolytic digested by Trypsin and prepared using Gel‐aided sample preparation [80]. The peptides were separated by LC (Ultimate3000 RSLC, Thermo) according to hydrophobicity in a reversed‐phase column (Acclaim PepMap, 300 µM ID—15 cm) and precolumn trapping. The MS/MS analyses were done in two steps as described [81]. Briefly, in the first step, an MS library is generated with pooled samples using the TOP30 method. In the second step, samples were analyzed by SWATH method. The data analyses were done using PeakView and MarkerView (both Sciex).

Asthma patients and healthy subjects

The asthma patients (n = 10) and healthy control cohort (n = 7) were recruited from the Department of Pulmonary Medicine, University Medical Center Essen‐Ruhrlandklinik, Essen, Germany. Three additional healthy control subjects were recruited at the Leibniz Research Centre for Working Environment and Human Factors (IfADo), Dortmund, Germany. The blood sampling from asthma patients was approved by the Westdeutsche Biobank Essen (University Medical Center Essen, approval number 22‐WBE‐142) and the sampling from healthy volunteers by the ethics committee of the Leibniz Research Centre for Working Environment and Human Factors (IfADo), Dortmund. Written informed consents for participation and publication were obtained. The blood samples were collected in heparin vials and analyzed on the same day. PBMCs were isolated and functionally checked by degranulation assay, IFN‐γ ELISA, and LC‐AA. Additionally, the PBMC compositions were determined by Cytek Aurora spectral flow cytometry using the following antibodies: CD3‐BUV563 (UCHT1, BD Biosciences), CD56‐BUV 805 (B159, BD Biosciences), CD226‐AF700 (DX11, BD Biosciences), CD314‐AF700 (FAB139N, R&D systems), CD335‐BV421 (9E2), CD244‐FITC (C1.7), CD336‐PerCP‐Cy5.5 (p44‐8), CD16‐PE Dazzle (3G8), CD337‐APC‐Fire750 (p30‐15), and Zombie NIR (all Biolegend). PBMCs were incubated with live/dead marker Zombie NIR for 15 min, washed, and subsequently stained with antibodies for 20 min at 4°C.

Statistics

Statistical analysis was performed using GraphPad Prism version 9.

Conflict of interest

The authors declare no financial or commercial conflicts of interest.

Author contributions

Martin Jürgens, Sabine Wingert, Maren Claus, Jens Alexander Niemann, Lea Katharina Picard, Elisabeth Hennes, and Jörg Reinders planned and carried out the experiments. Martin Jürgens, Sabine Wingert, Maren Claus, Birte Hellwig, Jörg Reinders, and Jörg Rahnenführer analyzed the data. Ina Haasler and Michaela Schedel provided essential reagents and support. Maren Claus, Silvia Capellino, and Carsten Watzl planned and supervised the project. Martin Jürgens and Carsten Watzl wrote the manuscript. Martin Jürgens, Maren Claus, Silvia Capellino, and Carsten Watzl edited the manuscript.

Consent

Written informed consents for participation and publication were obtained.

Peer review

The peer review history for this article is available at https://publons.com/publon/10.1002/eji.202451299.

Abbreviations

β2AR

β2‐adrenergic receptor

ECAR

extracellular acidification rate

GPCR

G protein‐coupled receptor

LABA

long‐acting β2‐agonist

LC‐AA

ligand complex adhesion assay

OCR

oxygen consumption rate

PKA

protein kinase A

Supporting information

Supporting Information

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.