In addition to their well-studied bronchodilatory and cardiotonic effects, β-adrenergic agonists carry anti-inflammatory properties by inhibiting cytokine production by human mononuclear cells. In a model of human promonocytic THP-1 cells stimulated with lipopolysaccharide (LPS), we showed that β-agonists inhibited tumor necrosis factor-α and interleukin-8 production predominantly via the β2-adrenergic receptor through the generation of cAMP and activation of protein kinase A. This effect was reproduced by other cAMP-elevating agents such as prostaglandins and cAMP analogs. Activation and nuclear translocation of the transcription factor nuclear factor-κB induced by LPS were inhibited with treatment with β-agonists, an effect that was prominent at late time points (>1 h). Although the initial IκB-α degradation induced by LPS was minimally affected by β-agonists, the latter induced a marked rebound of the cytosolic IκB-α levels at later time points (>1 h), accompanied by an increased IκB-α cytoplasmic half-life. This potentially accounts for the observed nuclear factor-κB sequestration in the cytoplasmic compartment. We postulate that the anti-inflammatory effects of β-agonists reside in their capacity to increase cytoplasmic concentrations of IκB-α, possibly by decreasing its degradation.
- nuclear factor-κB
- adenosine 3′,5′-cyclic monophosphate
β-adrenergic agonists are utilized in a variety of clinical situations, mostly for their bronchodilating and cardiotonic effects. It has been recognized that these pharmacological agents modulate the production of inflammatory mediators. It was, for example, shown that catecholamines could inhibit tumor necrosis factor (TNF)-α, interleukin (IL)-1, and IL-6 production by human mononuclear cells (16, 35, 42). It has been postulated that the generation of cAMP by these agents was necessary and that the regulation of this effect was at the level of inflammatory gene transcription (37-39).
Nuclear factor (NF)-κB is a transcription factor that has been implicated in control of the expression of numerous inflammatory genes including TNF-α and IL-8 (7). NF-κB is sequestered in an inactive form in the cytoplasm by its natural inhibitor, IκB-α. IκB-α phosphorylation and degradation induces the nuclear translocation of NF-κB and the binding of the protein to specific responding elements in the promoter regions of inflammatory genes (6). Several studies (1, 3, 33, 47) indicated that the regulation of the activation of this transcription factor is implicated in the anti-inflammatory or immunosuppressive effects of agents such as glucocorticoids, transforming growth factor-β1, and aspirin. We have therefore investigated whether the NF-κB pathway is implicated in the anti-inflammatory effects induced by β-agonists. In a model of human monocytic cells (THP-1 cells) stimulated with endotoxin, we found that the β2-adrenergic receptor was primarily implicated in the inhibition of IL-8 production observed with β-agonists. This inhibitory effect was cAMP- and protein kinase (PK) A-dependent and resided in the capacity of β-agonists to block the NF-κB pathway. We show here evidence that β-agonists modulate lipopolysaccharide (LPS) responses by inducing a marked increase of cytoplasmic IκB-α concentration in the presence of LPS, an effect that was observed only at late time points (>1 h).
MATERIALS AND METHODS
Cell culture and stimulation.
Human promonocytic THP-1 cells (American Type Culture Collection) were maintained in RPMI 1640 medium containing 10% fetal bovine serum (FBS), 2 mM l-glutamine, and antibiotics (all from Life Technologies, Paisley, UK). THP-1 cells were differentiated for 3 days with 10−7 M 1,25-dihydroxyvitamin D3 (Hoffmann La Roche, Basel, Switzerland) (43). The cells were washed and distributed into sterile microtiter plates (Costar, Corning, NY) at a concentration of 100,000 cells/well in RPMI 1640 medium containing 2% FBS. In some experiments, 1,25-dihydroxyvitamin D3-differentiated THP-1 cells were substituted with undifferentiated THP-1 cells transfected with CD14 (31). The cells were stimulated with nanomolar concentrations of Escherichia coli 0111:B4 LPS (List, Campbell, CA) for 8 h (unless indicated otherwise) at 37°C in the presence and absence of β-adrenergic agonists and antagonists. The following pharmacological agents were tested: isoproterenol (IMS, South El Monte, CA), albuterol (Glaxo Wellcome, Stevenage, UK), epinephrine (Sintetica, Mendrisio, Switzerland), norepinephrine (Aventis, Frankfurt, Germany), propranolol (Zeneca, Blackley, UK), metoprolol (Novartis, Basel, Switzerland), esmolol (Gensia, Bracknel, UK), phentolamine (Novartis), and iloprost (Ilomedin, Schering, Berlin, Germany). In most experiments, isoproterenol, a β1- and β2-adrenergic agonist, was used as a prototypic β-adrenergic agonist to inhibit IL-8 production in THP-1 cells. IL-8 and TNF-α concentrations were measured in conditioned supernatants with a sandwich ELISA with paired monoclonal antibodies available commercially (Endogen, Cambridge, MA) as described elsewhere (29).
In some experiments, the effect of β-agonists on both intracellular and secreted IL-8 production was assessed. To measure intracellular concentrations of IL-8, 5 × 106 cells were pelleted after 15 h of incubation and washed once with PBS, pH 2.0, to remove membrane-associated IL-8. The cells were then lysed in 20 mM Tris buffer, pH 7.5, containing 1% Triton X-100, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, 1 mg/ml of leupeptin, and 200 U/ml of aprotinin (Sigma, St. Louis, MO). Measurement of IL-8 concentrations was determined in parallel in conditioned supernatants and cell lysates.
Preparation of nuclear extracts and electrophoretic mobility shift assay.
1,25-Dihydroxyvitamin D3-differentiated THP-1 cells (7.5 × 106) were stimulated with 10 ng/ml of LPS in RPMI 1640 medium with 2% FBS for various times. After stimulation, the cells were rapidly chilled on ice, washed twice with ice-cold PBS, pH 7.4. Nuclear extracts were prepared as described elsewhere (28). Nuclear proteins were used for electrophoretic mobility shift assay. Twenty to fifty femtomoles of32P-labeled NF-κB double-stranded oligonucleotide probe (30,000–50,000 counts/min; 5′-AGT TGA GGG GAC TTT CCC AGG-3′; Promega, Madison, WI) were added to the nuclear proteins (5–8 μg) in a binding buffer containing 5 mM HEPES, pH 8.5, 5 mM MgCl2, 50 mM dithiothreitol, 0.4 mg/ml of poly(dI-dC) (Amersham Pharmacia Biotech, Uppsala, Sweden), 0.1 mg/ml of sonicated double-stranded salmon sperm DNA (Sigma), and 10% glycerol and incubated for 10 min at room temperature. Samples were migrated on a nondenaturing 5% acrylamide gel made in Tris-glycine-EDTA buffer. Gels were transferred onto Whatman paper, dried, and subjected to autoradiography.
Detection of IκB-α protein by Western blot.
After LPS stimulation in the presence and absence of β-agonists, THP-1 cells were lysed in 10 mM HEPES, pH 7.9, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride. In some experiments, the cells were pretreated for 30 min with the PKA inhibitor H-89. In one experiment, isoproterenol was substituted for PGE2, another cAMP-increasing agent. Ten micrograms of cytoplasmic protein extracts were separated by SDS-PAGE (10% acrylamide-bis-acrylamide gel) and electrotransferred onto a nitrocellulose membrane (Bio-Rad, Hercules, CA). IκB-α was detected with a rabbit polyclonal anti-human IκB-α antibody (Santa Cruz Biotechnology, Santa Cruz, CA), a mouse anti-rabbit horseradish peroxidase-conjugated secondary antibody (Jackson ImmunoResearch, West Grove, PA), and enhanced chemiluminescence (Amersham Pharmacia). In one experiment, the cells were treated with 10 μg/ml of cycloheximide (Sigma) 3 h after LPS treatment (32). Nuclear extracts were then prepared after various incubation times as described in Preparation of nuclear extracts and electrophoretic mobility shift assay. Quantification of IκB-α levels was done with densitometry of the IκB-α bands with a Molecular Dynamics densitometer and ImageQuant software (Sunnyvale, CA).
We used 1,25-dihydroxyvitamin D3-differentiated THP-1 cells stimulated by bacterial endotoxin as a model for monocyte/macrophage activation (4, 30, 31). The concentration-dependent effects of β-agonists on LPS-induced IL-8 production by differentiated THP-1 cells are shown in Fig.1 A. Results are expressed as a percentage of inhibition relative to the activation induced by LPS alone. All β-agonists inhibited production of IL-8 at concentrations as low as 10−9 M and showed similar affinities. Fenoterol, albuterol, isoproterenol, and epinephrine had EC50 values of 3, 9, 13, and 15 nM, respectively. Norepinephrine was slightly less potent than the other agonists, with an EC50 value of 38 nM. Fenoterol, epinephrine, isoproterenol, and norepinephrine showed similar efficacies, with maximal inhibition levels of 65, 68, 71, and 73%, respectively. The efficacy of albuterol was found to be consistently less than that of other β-agonists in inhibiting LPS-induced IL-8 production, with a maximal inhibition level of ∼ 25%. This may reflect a partial agonistic activity of this compound as previously reported by others (9, 14). Similar results were obtained with β-agonists in undifferentiated THP-1 cells transfected with CD14 (IL-8 production). The pharmacological activity of isoproterenol was independent of the LPS concentrations used to activate THP-1 cells (tested from 1 ng/ml to 10 μg/ml). This effect was not restricted to the production of IL-8. We found that TNF-α production by THP-1 cells was inhibited to a similar extent (Fig.1 B). CD14 surface expression, determined by fluorescence-activated cell sorter analysis, was not influenced by treatment with LPS and β-agonists (data not shown). In all experiments described in materials and methods, the cells were found fully viable by the cell viability MTT assay (12).
We next pharmacologically characterized the β-adrenergic receptor involved in these effects by measuring the potencies of several β-adrenergic antagonists (Fig. 2). The results are expressed as a percentage of the LPS response. Isoproterenol decreased the response of LPS by 69%. The addition of ≤10−7 M metoprolol, a selective β1-antagonist, practically unaltered the effect of isoproterenol. When added at concentrations ≥ 10−6M, the LPS response was restored. Similar results were obtained with another specific β1-antagonist, esmolol (data not shown). ICI-118551, a selective β2-antagonist, and propranolol, a nonselective β1- and β2-antagonist, showed the highest potency to interfere with the agonistic (anti-inflammatory) activity of isoproterenol. ICI-118551 and propranolol at concentrations as low as 10−8 M partially inhibited and at concentrations ≥ 10−6 M completely inhibited the effects of isoproterenol. Both were markedly more potent than the β1-antagonist metoprolol. An α-antagonist, phentolamine, did not modify the β-agonistic effect of isoproterenol (data not shown).
To determine whether the inhibitory effect of isoproterenol was cAMP dependent, we studied the effect of other cAMP-elevating agents, a cAMP analog, and a PKA inhibitor. PGE2 and iloprost (a prostacyclin analog) are two cAMP-elevating agents in monocytes. Both inhibited IL-8 production induced by 10 ng/ml of LPS, with potencies of 0.3 and 0.9 nM for PGE2 and iloprost, respectively. However, these mediators showed lower efficacies than the β-agonists (PGE2, 43%; iloprost, 51%). The plasma membrane-permeable cAMP analog dibutyryl cAMP decreased LPS-induced IL-8 release to the levels observed with isoproterenol. A concentration-dependent effect of dibutyryl cAMP was observed, with a maximal inhibitory response of 60% obtained with a 0.1 mM concentration. The PKA inhibitor H-89 abrogated the pharmacological activity of isoproterenol at a 10 μM concentration. Together, these results indicate that isoproterenol acted as an “anti-inflammatory” agent principally through its interaction with a β2-adrenergic receptor at the surface of THP-1 cells, leading to the generation of cAMP and activation of PKA.
Because inhibitory effects of β-agonists could be at the level of protein secretion per se (44), we measured the intra- and extracellular concentrations of IL-8 in THP-1 cells stimulated with LPS in the presence and absence of β-agonists. IL-8 production was increased in both intra- and extracellular compartments in response to LPS. Isoproterenol blocked both intra- and extracellular IL-8 to similar levels, an effect that could be reversed by propranolol (Table1). These experiments suggest that the inhibitory effect of isoproterenol takes place before the secretion process.
We next addressed whether isoproterenol influenced the activation and nuclear translocation of NF-κB because this transcription factor has been shown to be of importance for the activation of proinflammatory mediators such as IL-8 and TNF-α (20, 36). NF-κB was found to be activated in nuclear extracts from THP-1 cells treated with LPS but not in those cells treated with medium only (Fig.3). The NF-κB-oligonucleotide bands from electrophoretic mobility shift assay gels were quantified by densitometry in three separate experiments. The addition of isoproterenol did not modify NF-κB activation by LPS at 20 and 60 min but significantly decreased the NF-κB signal at 120 and 180 min (results for 20 and 180 min shown in Fig. 3). The loading of the nuclear protein-oligoprobe mixture was similar in all lanes (data not shown).
We next postulated that the inhibition of NF-κB activation and nuclear translocation induced by β-agonists was at the level of the regulation by its natural inhibitor IκB-α. Figure4 A shows, with a Western blot technique on cytosolic THP-1 cell extracts, that treatment with LPS resulted in the early disappearance (5–30 min) of IκB-α due to its degradation. Repeated experiments (n = 3) indicated that isoproterenol induced a significantly more rapid degradation of IκB-α (results for 1 representative experiment shown in Fig.4 A). This was followed by the reappearance of IκB-α after ∼1 h due to its resynthesis. Isoproterenol dramatically and significantly increased the cytoplasmic concentration of IκB-α at late time points (>1 h; P < 0.01; 1 representative experiment shown in Fig. 4 B and pooled results from 5 independent experiments shown in Fig. 4 C). This effect was particularly marked at 3 h. This effect was not observed in the absence of LPS treatment; i.e., isoproterenol did not increase the cytoplasmic IκB-α levels by itself (1 representative experiment shown in Fig. 4 D). This was confirmed in three separate experiments (data not shown). Because modulation of the cAMP-PKA pathway modified IL-8 secretion in THP-1 cells treated with LPS, we next addressed whether another cAMP-increasing agent, PGE2, could also influence IκB-α cytosolic concentrations. As shown in Fig. 5 A, PGE2 had a similar inhibitory effect compared with isoproterenol. This was a consistent finding observed in four separate experiments (Fig.5 B). Increased IκB-α cytoplasmic levels induced by isoproterenol at 150 min could be blocked by the addition of the PKA inhibitor H-89 (Fig. 5 B). These results strongly suggested that this effect was not specific for β-agonists but was observed with other cAMP-increasing agents such as PGE2 and that a cAMP-PKA-dependent pathway was directly involved in the modulation of IκB-α in response to these pharmacological agents.
The marked increase in cytoplasmic levels of IκB-α protein induced by isoproterenol may result from an increase in IκB-α synthesis or a decrease in its degradation. To address this, we performed an experiment in which we tested the stability of the IκB-α protein in the cytoplasm after stimulation with LPS in the presence and absence of isoproterenol. The cells were stimulated for 3 h with LPS in the presence and absence of isoproterenol. The protein synthesis inhibitor cycloheximide (10 μg/ml) was then added, and cytoplasmic IκB-α levels were measured with Western blots at different times over the next 2 h. Figure 6 shows that after LPS stimulation, IκB-α levels remained high, with greater stability in the presence of the β-agonist than that measured in its absence (half life ∼60 and 20 min, respectively).
Here we show that β-adrenergic agonists inhibit the production of IL-8 by human promonocytic THP-1 cells in response to LPS. The inhibitory effect of β-adrenergic agonists on LPS-induced IL-8 production was predominantly mediated by their interaction with the β2-adrenergic receptor. This conclusion was based on experiments with β-agonists of different selectivities for the β1- and β2-adrenergic receptors and the use of selective and nonselective β-blockers. We found that albuterol and fenoterol, selective β2-agonists, had similar potencies compared with epinephrine and isoproterenol in inhibiting LPS-dependent mononuclear cell activation. In contrast, norepinephrine, known for its greater avidity for the β1-adrenergic receptor, was three times less potent than the other agonists tested. In addition, the compound ICI-118551, a selective β2-blocker, and propranolol, a nonselective β1- and β2-antagonist, were much more efficient in reverting the β-agonistic effect than the selective β1-agonist metoprolol. This is in accordance with the work of Sekut et al. (34). These authors reported the inhibitory effect of the β2-selective agonists albuterol and salmeterol on LPS-induced TNF-α production by these cells, an effect that was blocked by the specific β2-antagonist oxprenolol (34). In addition, the β2-adrenoceptor was previously found to be the adrenoceptor prominently expressed and functional in monocyte/macrophage-like cells (8, 17, 18, 21,25). In contrast, Talmadge et al. (39) found that in undifferentiated promonocytic THP-1 cells, the β1-adrenoceptor mediated the inhibition of LPS-induced TNF synthesis (39). This is probably due to a different β1- to β2-adrenoceptor ratio of surface expression in undifferentiated THP-1 cells as well as a different readout (mRNA vs. protein). The cell type utilized in our study (macrophages) as well as the agonist and antagonist potency orders indicate that participation of the β3-adrenoceptor is unlikely.
The observed decreased IL-8 production by β-agonists was dependent on the generation of cAMP and on the activation of PKA. Indeed, the cAMP analog dibutyryl cAMP and other cAMP-elevating agents such as PGE2 and iloprost reproduced the effects observed with β-agonists. H-89, a PKA inhibitor, completely reversed the anti-inflammatory effects of isoproterenol. These findings are in agreement with several studies (2, 10, 37, 38) that tested the effects of cAMP-elevating agents. Whether downstream effectors of this pathway [e.g., cAMP response element binding protein (CREB) transcription factor] are implicated remains to be determined.
Because NF-κB has been implicated in the transcriptional regulation of many inflammatory genes including TNF-α (46) and IL-8 (13, 20), we addressed whether β-agonists would interfere with this pathway. We found that this effect was likely due to decreased activation and nuclear translocation of NF-κB, and this was observed only after >1 h of LPS stimulation. At earlier time points, the drug did not influence the level of NF-κB activation. We hypothesized that this was due to a secondary increase of cytoplasmic concentrations of its natural inhibitor IκB-α, a mode of action already described for other anti-inflammatory agents (3,33). Cytoplasmic levels of IκB-α as measured by Western blot indicated that the treatment of cells with isoproterenol did not prevent the initial IκB-α degradation on LPS stimulation but rather induced a subsequent marked increase in cytosolic IκB-α levels. Even a small increase in cytosolic concentrations of IκB-α was previously shown to negatively affect NF-κB nuclear translocation (19). Importantly, and in contrast with the mode of action described for glucocorticoids, an increased production of IκB-α was not observed in cells treated only with β-agonists. The addition of a proinflammatory stimulus such as LPS was necessary to observe this effect.
Ollivier et al. (24) previously showed that cAMP induced by forskolin inhibited LPS-induced NF-κB activation in THP-1 cells. In their study, the rate of NF-κB nuclear translocation was not affected, but it was the transcription efficiency of NF-κB at early times (1 h) that was reduced with increased cAMP concentrations. However, these authors did not investigate the activation of NF-κB at later time points. Our results unravel another possible mechanism, which may explain the inhibitory effect observed with β-agonists. After the initial NF-κB activation by LPS, β-agonists may induce the activation of transcription factors such as CREB, which might cooperate with NF-κB at the level of the IκB-α promoter to increase its transcription. Such a cooperative mechanism has been described for NF-κB and other transcription factors (20). This could be the reason for the need of the presence of both LPS and β-agonists to observe the increased IκB-α cytoplasmic levels. Interestingly, recently published data (11) indicated that a transcriptionally active glucocorticoid receptor was translocated into the nucleus on treatment with β2-agonists, which may also cooperate with NF-κB to increase IκB-α transcription.
Another possibility is that the regulation of IκB-α by cAMP is at the level of protein degradation as suggested by the experiment shown in Fig. 5. Such an effect was proposed by Neumann et al. (22) in a study where forskolin increased the cytoplasmic levels of IκB-α. It is also conceivable that β-agonists decrease IκB-α phosphorylation and degradation through the activation of second messengers, which will, in turn, inhibit upstream kinases such as IL-1 receptor-associated kinase, TNF receptor-associated kinase-6, or members of the mitogen-activated protein kinase kinase kinase (MEKK)-1-NF-κB-inducing kinase-IκB kinase complex involved in LPS signaling (15, 23). The delayed appearance of IκB-α protein in the cytoplasm could also suggest that β-agonists may induce or activate a cytosolic inhibitor of kinases upstream of NF-κB. Candidate inhibitors are those of the family of antiapoptotic factors. Indeed, it was recently demonstrated that increased expression of Bcl-2 and Bcl-XL, which are under the control of both NF-κB and CREB (40, 45), prevented IκB-α degradation (5). In another study, it was shown that the LPS-induced A20 protein could directly inhibit MEKK-1 (Kravchenko VV, personal communication). It is also possible that an anti-inflammatory cytokine such as IL-10 is produced on β-agonist treatment, which may turn down IL-8 production induced by LPS in an autocrine fashion via an IκB-dependent mechanism (27,41). Finally, Parry and Mackman (26) have proposed that NF-κB inhibition by cAMP occurred at the level of the differential binding of CREB and NF-κB to the CREB-binding protein, a protein necessary for efficient gene transcription (26).
In conclusion, we hereby provide clues as to the mechanisms by which cAMP-increasing agents such as β-agonists exert their anti-inflammatory effects. These pharmacological agents block NF-κB activation and nuclear translocation and, secondarily, inflammatory gene transcription by increasing IκB-α cytoplasmic concentration. Our results also make an important link between the cAMP and NF-κB pathways.
We thank Pierre Weber (Hoffmann La Roche, Basel, Switzerland) and Ursula Lang and Alessandro Capponi (University of Geneva, Geneva, Switzerland) for the gift of precious reagents, and A. Nials, M. Skingle and T. N. C. Wells for stimulating discussions and constant support.
This work was supported by Swiss National Foundation for Scientific Research Grant SNF 32-50764 (to J. Pugin) and grants from the 3R and Carlos and Elise de Reuter Foundations and Glaxo Wellcome.
P. Farmer received a scholarship from the Canadian Heart and Stroke Foundation and the Fonds pour la Formation de Chercheur et l'Aideà la Recherche. J. Pugin is the recipient of a fellowship from the Prof. Dr. Max Cloëtta Foundation.
Address for reprint requests and other correspondence: J. Pugin, Division of Medical Intensive Care, Dept. of Internal Medicine, University Hospital of Geneva, 1211 Geneva 14, Switzerland (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2000 the American Physiological Society