Abstract
Chronic use of β2-adrenoceptor agonists as a monotherapy in asthma is associated with a loss of disease control and an increased risk of mortality. Herein, we tested the hypothesis that β2-adrenoceptor agonists, including formoterol, promote biased, β-arrestin (Arr) 2–dependent activation of the mitogen-activated protein kinases, ERK1/2, in human airway epithelial cells and, thereby, effect changes in gene expression that could contribute to their adverse clinical outcomes. Three airway epithelial cell models were used: the BEAS-2B cell line, human primary bronchial epithelial cells (HBEC) grown in submersion culture, and HBEC that were highly differentiated at an air-liquid interface. Unexpectedly, treatment of all epithelial cell models with formoterol decreased basal ERK1/2 phosphorylation. This was mediated by cAMP-dependent protein kinase and involved the inactivation of C-rapidly-activated fibrosarcoma, which attenuated downstream ERK1/2 activity, and the induction of dual-specificity phosphatase 1. Formoterol also inhibited the basal expression of early growth response-1, an ERK1/2-regulated gene that controls cell growth and repair in the airways. Neither carvedilol, a β2-adrenoceptor agonist biased toward βArr2, nor formoterol promoted ERK1/2 phosphorylation in BEAS-2B cells, although β2-adrenoceptor desensitization was compromised in ARRB2-deficient cells. Collectively, these results contest the hypothesis that formoterol activates ERK1/2 in airway epithelia by nucleating a βArr2 signaling complex; instead, they indicate that β2-adrenoceptor agonists inhibit constitutive ERK1/2 activity in a cAMP-dependent manner. These findings are the antithesis of results obtained using acutely challenged native and engineered HEK293 cells, which have been used extensively to study mechanisms of ERK1/2 activation, and highlight the cell type dependence of β2-adrenoceptor–mediated signaling.
SIGNIFICANCE STATEMENT It has been proposed that the adverse effects of β2-adrenoceptor agonist monotherapy in asthma are mediated by genomic mechanisms that occur principally in airway epithelial cells and are the result of β-arrestin 2–dependent activation of ERK1/2. This study shows that β2-adrenoceptor agonists, paradoxically, reduced ERK1/2 phosphorylation in airway epithelia by disrupting upstream rat sarcoma–C-rapidly accelerated fibrosarcoma complex formation and inducing dual-specificity phosphatase 1. Moreover, these effects were cAMP-dependent protein kinase–dependent, suggesting that β2-adrenoceptor agonists were not biased toward β-arrestin 2 and acted via canonical, cAMP-dependent signaling.
Introduction
Inhaled β2-adrenoceptor agonists are a mainstay asthma therapy and effectively relieve symptoms by promoting rapid and prolonged bronchodilation. However, regular administration of these drugs as a monotherapy is associated with tolerance (Cockcroft et al., 1993; Salpeter et al., 2006), a loss of asthma control (Drazen et al., 1996), and increased mortality (Nelson et al., 2006). The mechanism driving these undesirable clinical outcomes remains unclear, but the ability of β2-adrenoceptor agonists to modulate the expression of numerous proinflammatory genes may be an important factor (Ritchie et al., 2018; Yan et al., 2018). Indeed, β2-adrenoceptor agonists, administered chronically to sensitized mice, enhance the “asthma-like” pathologic changes that follow allergen challenge, including airway hyper-responsiveness, mucus hypersecretion, pulmonary inflammation, and increased pulmonary leukocyte burden (Lin et al., 2012; Thanawala et al., 2013). Mechanistically, Nguyen et al., (2017) have proposed that β2-adrenoceptor agonists stimulate Gαs in airway smooth muscle to effect bronchodilation and relieve symptoms but exacerbate asthma pathology by activating mitogen-activated protein (MAP) kinase signaling in a β-arrestin (Arr) 2–dependent manner (vide infra). They also posit that this “biased agonism” is mediated by β2-adrenoceptors located on airway epithelial cells. This is a credible proposal because the airway epithelium promotes and regulates inflammatory processes, is the first site of action for inhaled therapies, and expresses a population of efficiently coupled β2-adrenoceptors (Davis et al., 1990; Penn et al., 1994; Kelsen et al., 1995; Knight and Holgate, 2003; Joshi et al., 2021). Moreover, extracellular signal-regulated kinases (ERK) 1 and 2 are phosphorylated to a greater degree in bronchial biopsies taken from subjects with asthma than in healthy controls, with changes in the airway epithelium being positively correlated with indices of inflammation (Liu et al., 2008; Alam and Gorska, 2011).
Classically, β2-adrenoceptor agonists promote the Gαs-dependent stimulation of adenylyl cyclase, cAMP generation, and activation of several downstream effectors (Giembycz and Newton, 2006; Gerits et al., 2008). One target is cAMP-dependent protein kinase (PKA), which regulates MAP kinase signaling in a complex manner (Dumaz and Marais, 2005; Gerits et al., 2008). In some tissues, the MAP kinase kinase kinase B-rapidly accelerated fibrosarcoma (Raf) is activated indirectly by PKA via the small GTP-binding protein rat sarcoma (Ras)-proximate 1 and promotes the sequential phosphorylation of mitogen-activated protein kinase kinase (MEK) 1/2 and ERK1/2 (Vossler et al., 1997; Takahashi et al., 2017b). In other cell types, the B-Raf paralog, C-Raf, is phosphorylated by PKA and, thereby, rendered unable to interact with Ras, such that downstream MEK1/2-ERK1/2 signaling is impaired (Cook and McCormick, 1993). β2-Adrenoceptor agonists also modulate ERK1/2 activity in a cell type–dependent manner (Lefkowitz et al., 2002). For example, ERK1/2 is activated in HEK293 cells (Daaka et al., 1997; Schmitt and Stork, 2000; van der Westhuizen et al., 2014), human dermal fibroblasts (Pullar and Isseroff, 2006), rat cardiac myocytes (Zou et al., 1999), COS-7 fibroblasts (Crespo et al., 1995), and rat dorsal root ganglion neurons (Aley et al., 2001) but inactivated in the MDA-MB-231 human breast cancer cell line (Pon et al., 2016) and J774 murine macrophages (Keränen et al., 2017). Currently, the cell type–specific factors that determine whether PKA increases or decreases MAP kinase signaling are unclear.
β2-Adrenoceptor agonism also leads to the recruitment of β-arrestins. These multifunctional proteins uncouple the agonist-occupied receptor from Gs, resulting in homologous desensitization (Benovic et al., 1987; Goodman et al., 1996; Luttrell and Lefkowitz, 2002). They also act as scaffolds for the assembly of a myriad of signaling elements that include C-Raf, MEK1/2, and ERK1/2 (Peterson and Luttrell, 2017). Currently, the role of β-arrestins in β2-adrenoceptor–mediated ERK1/2 activation is controversial. Although abundant data suggest that β-arrestins are obligatory (Shenoy et al., 2006; Luttrell and Gesty-Palmer, 2010; Luttrell et al., 2018), the application of genome editing technologies has recently challenged that assumption (O'Hayre et al., 2017; Grundmann et al., 2018).
In this study, we tested the hypothesis that β2-adrenoceptor agonists activate βArr2–dependent C-Raf-MEK1/2-ERK1/2 signaling in airway epithelial cells and, thereby, promote changes in gene expression that could contribute to their adverse effects in asthma. Three airway epithelial cell models were compared in this endeavor: the BEAS-2B bronchial epithelial cell line, human primary bronchial epithelial cells (HBEC) grown in submersion culture, and HBEC that were highly differentiated at an air-liquid interface (ALI). HEK293 cells were examined in parallel, as they have been used extensively to interrogate the impact of β2-adrenoceptor agonism on MAP kinase signaling and are an instructive comparator.
Materials and Methods
Submersion Culture of HEK293β2 Cells
HEK293S cells stably expressing an amino-terminal, hemagglutinin-tagged human β2-adrenoceptor (HEK293β2) at a density of 3.2 pmol/mg protein were donated by Dr. Michel Bouvier (University of Montreal, PQ, Canada). Cells were grown until confluent under a 5% CO2/air atmosphere at 37°C in Dulbecco’s modified Eagle’s medium supplemented with 10% FBS, l-glutamine (2 mM), and G-418 (400 μg/ml). Cells were cultured for a further 24 hours in supplement-free basal medium prior to experimentation.
Submersion Culture of BEAS-2B Cells
BEAS-2B cells (ATCC, Manassas, VA) were cultured for 2 days under a 5% CO2/air atmosphere at 37°C in 24-well tissue culture plates (Costar Inc, Corning, NY) containing Dulbecco’s modified Eagle’s medium/F12 supplemented with 10% FBS (Life Technologies, Burlington, Ontario, Canada), l-glutamine (2.5 mM), and NaHCO3 (14 mM; all Invitrogen) and for a further 24 hours in serum-free medium (SFM) without supplements (Greer et al., 2013). At this time, cultures were confluent and used for experimentation.
Submersion Culture of HBEC
Ethics approval for the use of human tissues was granted by the Conjoint Health Research Ethics Board of the University of Calgary. Bronchial epithelial cells were obtained by proteinase digestion of nontransplanted, normal airways obtained from a tissue retrieval service at the International Institute for the Advancement of Medicine (Edison, NJ).
HBEC grown as monolayers in submersion culture were seeded in 12-well plates (Costar) containing bronchial epithelial cell growth medium (PromoCell, Heidelberg, Germany) supplement penicillin (50 μg/ml) and streptomycin (10 μg/ml) and maintained for ∼14 days under a 5% CO2/air atmosphere at 37°C until ∼80% confluent. Cells were cultured for a further 24 hours in supplement-free basal medium (PromoCell) prior to experimentation.
Culture of HBEC at an ALI
Highly differentiated human airway epithelial cell cultures were generated as described in detail previously (Michi and Proud, 2021). Briefly, HBEC were expanded in T75 cm2 flasks (Costar) and maintained for 72 hours at 37°C under a 5% CO2/air atmosphere in PneumaCult-EX expansion medium (Stemcell Technologies, Vancouver, BC, Canada). This was replaced every 48 hours until cells were 95% confluent, at which time they were dissociated (TrypLE Select, Invitrogen) and resuspended in F12 medium containing 20% FBS and pelleted by centrifugation. Cells were resuspended in PneumaCult-EX expansion medium and seeded at a density of 2 × 105/cm2 in 0.4-μm pore transwell inserts (Costar) coated with bovine collagen type I/III (Advanced BioMatrix, San Diego, CA). At 48 hours, the PneumaCult-EX expansion medium was replaced with PneumaCult-ALI differentiation medium (Stemcell Technologies) containing 10× supplement, fluconazole (25 μg/ml), and penicillin and streptomycin (each 10 μg/ml). Cultures were fed every 48 hours for 5 weeks with PneumaCult-ALI differentiation medium containing 100× supplement, hydrocortisone (0.5 μg/ml), and heparin (50 μg/ml). At 14 days post-transwell seeding, cells were washed apically once per week with PBS to remove mucus. Prior to experimentation, ALI cultures were maintained for 18 hours in PneumaCult basal medium (Stemcell Technologies), which lacks all supplements, including hydrocortisone, and washed with PBS. Drugs were administered simultaneously to both the apical and basolateral surfaces.
Stable Generation of a CRE Reporter in BEAS-2B Cells and Measurement of Luciferase Activity
BEAS-2B cells were transfected with plasmid DNA (pADneo2-C6-BGL) to generate 6×CRE luciferase reporter cells as described previously (Meja et al., 2004). Luciferase activity was measured by luminometry and expressed as a fold change relative to time-matched, vehicle-treated cells (Yan et al., 2018).
Expression of a Protein Inhibitor of PKA in HBEC and BEAS-2B Cells
Subconfluent (70%) HBEC and BEAS-2B cells were infected with an adenovirus vector (Ad5.CMV.PKIα) containing a DNA fragment encoding the amino acid sequence of the cAMP-dependent protein kinase inhibitor (PKI) protein α (Meja et al., 2004). A vector (Ad5.CMV.GFP) encoding GFP was used to control for any off-target effects of the virus, per se. Cells were cultured for 48 hours at 37°C as described above and for a further 24 hours in SFM. Immunofluorescence confocal microscopy and Western blotting were employed to confirm expression of the transgenes.
Derivation of Clonal BEAS-2B Cells Deficient in βArr2
Cells were subjected to CRISPR/Cas9 genome editing using a proprietary gene knockout kit (12501; Synthego, Redwood City, CA) according to the manufacturer’s protocol. Briefly, cells were grown to 50%–70% confluence in 24-well plates. A chemically modified single-guide RNA (180 pmol) encoding a sequence (5′-CGTGAAGACCTGGATGTGCTGGG-3′) that targets ARRB2 was mixed with Cas9 nuclease (20 pmol from Streptococcus pyogenes) in OptiMEM media (31985070; Thermo Scientific, Waltham, MA) containing 1 μl Lipofectamine Cas9 Plus reagent (CMAX00008; Thermo Scientific). The mixture was incubated at room temperature for 10 minutes, added to OptiMEM medium containing 1.5 μl LipofectAMINE CRISPRMAX transfection reagent (Thermo), and transfected into BEAS-2B cells in the presence of serum. At 48 hours, cells were seeded into 96-well plates, and single clones were isolated by limited dilution. Clones were cultured for 10–14 days and subjected to Western blotting to confirm the absence of βArr2. Three positive clones (C1, C2, C3) were identified, and the deletion of ARRB2 was verified by PCR genotyping using primers (forward: 5′-GCTAGGGAAGTGAAATGGGC-3′; reverse: 5′-TCACGGTGAAGAAGAAGGGG-3′) that flanked both deleted and undeleted regions. PCR products were subcloned into the pEGFP-C2 vector (Takara Bio, Mountain View, CA) at the EcoRI and KpnI sites; Sanger sequencing of plasmid DNA was then performed at the Centre for Genome Engineering, University of Calgary. Supplemental Fig. 1 shows the sequencing of C1, which revealed the insertion of a thymine nucleotide resulting in a frameshift mutation and the generation of a premature stop codon.
Knockdown of DUSP1 and ARRB1
Parental BEAS-2B cells were grown to 60%–70% confluence in 12-well plates before being transfected with siRNA s. Two DUSP1-targeting siRNAs (SI00374801: 5′-TAGCGTCAAGACATTTGCTGA-3′; SI00374808: 5′-CTGTACTATCCTGTAAATATA-3′) or a nontargeting siRNA control (SI03650325: 5′-AATTCTCCGAACGTGTCACGT-3′) (all Qiagen) were mixed with 3 µl of Lipofectamine RNAiMax (13778150; Thermo Scientific) in 100 µl of OptiMEM and incubated at room temperature for 5 minutes. Cells were cultured for 24 hours in the presence of Dulbecco’s modified Eagle’s medium/F12 supplemented with 10% FBS containing each siRNA at a final concentration of 25 nM as indicated. Cells were then incubated for an additional 24 hours in SFM prior to experimentation. In βArr1 knockdown experiments, pools of four ARRB1-targeting siRNAs (SI02776921: 5′-CTCGACGTTCTGCAAGGTCTA-3′; SI02643977: 5′-CGGTGTGGACTATGAAGTCAA-3′; SI00058961: 5′-CACCAACAAGACGGTGAAGAA-3′; SI00058954: 5′-TACCA ATCTCATAGAACTTGA-3′) or nontargeting control siRNAs (SI03650325: 5′-AATTCTCCGAACGTGTCACGT-3′; SI04380467: 5′-AAGCAGCACGACTTCTTCAAG-3′; SI1022064: 5′-CGGCAAGCTGACCCTGAAGTTCAT-3′; SI1027280 :) (proprietary; all Qiagen) were transfected into BEAS-2B or HEK293β2 cells using LipofectAMINE RNAiMax at final concentrations of 0.1, 1, or 10 nM.
RNA Extraction and Gene Expression
Total RNA was extracted using Nucleospin RNA mini kits (Macherey-Nagel Inc., GmbH & Co., Duren, Germany) and reverse-transcribed to cDNA using a qScript cDNA synthesis kit as described by the manufacturer (Quanta Biosciences, Gaithersburg, MD). The abundance of mRNA encoding DUSP1 and early growth response 1 (EGR1) was measured by real-time PCR using a StepOnePlus instrument (Applied Biosystems) as described previously (Joshi et al., 2017; Yan et al., 2018) and expressed either as a ratio to GAPDH or as a fold change relative to vehicle-treated cells as described. Primer sequences were as follows: DUSP1 (forward, 5′-CGCGCAAGTCTTCTTCCTCA-3′; reverse, 5′-GATGCTTCGCCTCTGCTTCA-3′), EGR1 (forward, 5′-ACCTGACCGCAGAGTTTTT-3′; reverse, 5′-GAGTGGTTTGGCTGGGGTAA-3′), and GAPDH (forward, 5′-ATGGAAATCCATCACCATCTT-3′; reverse, 5′-CAGCATCGCCCCACTTG-3′).
Western Immunoblot Analysis
Confluent BEAS-2B cells, HBEC, and HEK293β2 cells were lysed in 1× Laemmli sample buffer (4% SDS, 10% 2-mercaptoehtanol, 20% glycerol, 0.004% bromophenol blue, 125 mM Tris-HCl, pH 6.8) supplemented with phosphatase inhibitors (Sigma-Aldrich) and 1× complete protease inhibitor cocktail (Roche, Basal, Switzerland). Cell lysates were size-fractionated on SDS polyacrylamide gels, electrotransferred onto reinforced 0.2-µm nitrocellulose membranes (GE Healthcare, Waukesha, WI), and blocked with 5% milk in Tris-buffered saline containing 1% Tween 20. Membranes were probed with primary antibodies against total (t) ERK1/2 (cs-4695), phospho (p) ERK1/2 (cs-9101), tMEK1/2 (cs-9122), pMEK1/2 (cs-9154), C-Raf (cs-53745), pSer259-C-Raf (cs-9421), pSer338-C-Raf (cs-9427), βArr1 (cs-12697), βArr2 (cs-3857), EGR1 (cs-4153), GFP (cs-2555) (Cell Signaling Technology, Danvers, MA), DUSP1 (sc-1102), Elk -1 (sc-365876), pElk-1 (sc-8406), PKIα (sc-1943) (Santa Cruz Biotechnology, Dallas, TX), pSer43-C-Raf (ab-150365), pSer621-C-Raf (ab-157201) (Abcam, Cambridge, MA), or GAPDH (MCA4739, Bio-Rad, Hercules, CA) overnight at 4°C. Membranes were incubated with horseradish peroxidase–linked secondary immunoglobulin (Jackson ImmunoResearch Laboratories Inc., West Grove, PA) for 1 hour at room temperature and detected by enhanced chemiluminescence (Thermo Scientific). Images were acquired using a ChemiDoc Touch imaging system (Bio-Rad) and analyzed with ImageLab software (Bio-Rad). Band volumes of interest were normalized as indicated in each figure or presented as a fold change from unstimulated levels at time 0.
Measurement of ERK1/2 Phosphorylation by Electrochemiluminescence
Phospho-ERK1/2 (Thr202/Tyr204; Thr185/Tyr187) and total ERK1/2 in cell lysates prepared from confluent BEAS-2B cells, HBEC, or HEK293β2 cells treated with the drug of interest or vehicle were measured by electrochemiluminescence (ECL) using the MesoScale Discovery whole cell lysate multispot kit (K15107D-2) with a MESO QuickPlex SQ 120 instrument (MesoScale Discovery, Gaithersburg, MD).
Determination of cAMP Mass
Confluent wild-type and βArr2-deficient BEAS-2B cells at 37°C were treated with the drug of interest or vehicle. The culture medium was decanted, and cells were lysed with HCl (0.1M). After neutralization with NaOH, cAMP mass was measured by ECL using a 96-well assay kit (K150FDD-2, MesoScale Discovery) according to the manufacturer’s instructions and quantified using Discovery workbench 4.0 software (MesoScale Discovery). Total protein in each sample was measured using a BCA protein assay (23227; Thermo Scientific), and cAMP was expressed in units of picograms per microgram protein.
Gene Expression Profiling
HBEC in submersion culture were treated with formoterol (1 nM) or vehicle for 1, 2, 6, and 18 hours. The relative expression of DUSP isoforms and components of the βArr- and Gβγ-signaling pathways were determined by RNA-sequencing as described previously (Joshi et al., 2019) and are expressed in transcripts per million (TPM). These data are freely available via NCBI ’s Gene Expression Omnibus (accession code pending). The effects of formoterol (1 nM), salmeterol (100 nM), and indacaterol (100 nM) on gene expression changes in BEAS-2B cells were mined from previous microarray or RNA-sequencing data generated by the authors (accession codes: GSE106710, GSE115830, and GSE126981).
Curve Fitting
Monophasic, agonist concentration-response curves were fit by least-squares, nonlinear, iterative regression to the following equation (Prism 6; GraphPad Software Inc, San Diego, CA):
where E is the effect; Emin and Emax are the basal and maximum responses, respectively; [A] is the molar concentration of agonist; EC50 is the molar concentrations of agonist that produces (Emax − Emin)/2; and n is the Hill coefficient. Agonist potency is expressed as −log10 EC50 (pEC50).
The time-dependent phosphorylation and dephosphorylation of elements of the C-Raf signaling cascade in response to drug treatment were fit to one-phase exponents of the form
respectively, where x is time, E is the response at time x, E0 is the response when x = 0, k is the rate constant, and p is the plateau at infinite time; for dephosphorylation, p may be greater than, or equal to, a value of zero. The time taken for drug treatment to generate 50% of the maximal response (t1/2) is given by ln(2)/k.
Drugs and Analytical Reagents
R,S-salmeterol xinafoate, R,R-formoterol fumarate, and L-161,982 were donated by GlaxoSmithKline (Stevenage, UK), AstraZeneca (Mölndal, Sweden), and Merck Frosst, Inc. (Montreal, PQ, Canada), respectively. R,S-carvedilol hydrochloride, R,S-alprenolol hydrochloride, and ICI 118,551 were purchased from Tocris Bioscience (Toronto, ON, Canada). ONO-AE1-259 and ONO-AE1-329 were donated by ONO Pharmaceuticals (Osaka, Japan). R,S-carazolol hydrochloride and R,S-vilanterol trifenatate were from Toronto Research Chemicals (North York, ON, Canada). R-indacaterol maleate, roflumilast N-oxide (RNO), TG4-155, and YM-254890 were from Gilead Sciences (Seattle, WA), Altana (Konstanz, Germany), ChemDiv (San Diego, CA), and Focus Biomolecules (Plymouth, PA), respectively. Forskolin, R,S-atropine sulfate, mepyramine maleate, PD 098059, U0126, and phorbol 12-myristate 13-acetate (PMA) were obtained from Sigma-Aldrich (St. Louis, MO). All drugs were dissolved in DMSO and diluted to the required working concentration in aqueous medium. The final concentration of DMSO was ≤0.2% and did not affect any outcome measured. Histamine dihydrochloride, carbachol chloride, and G-418 were purchased from Sigma-Aldrich and dissolved in sterile water. Epidermal growth factor (EGF) and tumor necrosis factor-α (TNFα) were obtained from R&D Systems (Minneapolis, MN) and dissolved in PBS containing 0.1% bovine serum albumin .
Statistics
Data are presented as either the mean ± S.E.M. or, more typically, as box and whisker plots (showing all data points) of N independent measurements. Significant changes in luciferase reporter activity and gene expression were determined by using Student’s two-tailed, paired t test, or repeated measures one-way ANOVA, as indicated. When the ANOVA F-test P value was <0.05, differences between groups were determined by using Tukey’s multiple comparisons test without Greenhouse-Geisser correction. The time-dependent change in ERK1/2 phosphorylation between two treatments was by repeated measures two-way ANOVA followed by Sidak’s multiple comparisons test. The null hypothesis was rejected when P < 0.05.
Results
Formoterol and Salmeterol Promoted ERK1/2 Phosphorylation in HEK293β2 Cells
In agreement with several previous studies (Shenoy et al., 2006; van der Westhuizen et al., 2014; Luttrell et al., 2018), Western blotting determined that ERK1 and ERK2 were partially phosphorylated in unstimulated HEK293β2 cells (Fig. 1). The long-acting β2-adrenoceptor agonist (LABA) formoterol (1 nM) produced a transient increase in the phosphorylation of both ERK isoforms relative to vehicle-treated cells matched for time, which peaked at 5 minutes (∼8-fold increase) and then declined to a new steady state that at 60 minutes was ∼2-fold above the prestimulated level (Fig. 1A). This effect was abolished in cells pretreated (30 minutes) with ICI 118,551 (1 μM), a competitive and selective β2-adrenoceptor antagonist (Fig. 1B). The ability of formoterol to phosphorylate ERK1/2 was concentration-dependent, with an EC50 measured at 5 minutes of 5 pM (Fig. 1C). Similarly, treatment of cells with EGF (10 ng/ml; 30 minutes) produced a robust phosphorylation of ERK1/2 that was greater than that produced by formoterol (Fig. 1A). Comparable data were obtained in HEK293β2 cells treated with a structurally dissimilar LABA, salmeterol (Supplemental Fig. 2, A and B). Total ERK1/2 levels were unaffected by these interventions.
Formoterol and Salmeterol Promoted ERK1/2 Dephosphorylation in Airway Epithelial Cells
In unstimulated BEAS-2B cells and HBEC, Western blotting similarly detected a basal level of pERK1 and pERK2 (Fig. 2; Supplemental Figs. 3 and 4). However, in both epithelial cell models, formoterol (1 nM) produced a time-dependent and sustained dephosphorylation of these ERK isoforms (t1/2 ∼10 minutes) that had plateaued to a level that was at 10%–30% of the control at 60 minutes (Fig. 2, A and B). In cells pretreated (30 minutes) with ICI 118,551 (1 μM), ERK1/2 phosphorylation was abolished, indicating that this was a β2-adrenoceptor–mediated effect (Fig. 2, C and D). At the 60-minute time point, formoterol dephosphorylated ERK1/2 in BEAS-2B and HBEC in a concentration-dependent manner, with EC50 values of 12 pM and 22 pM, respectively (Fig. 2, E and F). Comparable data were produced with salmeterol (100 nM) and another LABA, indacaterol (100nM), indicating that ERK1/2 dephosphorylation was a class effect of β2-adrenoceptor agonists (Supplemental Figs. 3 and 4). In contrast, exposure of BEAS-2B cells and HBEC to EGF (10 ng/ml; 30 minutes), TNFα (10 ng/ml; 60 minutes), and PMA (100 nM; 30 minutes) enhanced ERK1/2 phosphorylation, indicating that the basal levels measured in unstimulated cells were not maximal (Fig. 2; Supplemental Figs. 3 and 4). Total ERK1/2 expression was unaffected by these interventions (Fig. 2; Supplemental Figs. 3 and 4).
The human airway epithelium has a pseudostratified columnar architecture, which largely influences its mucosal defense properties and function. HBEC isolated from the main stem bronchi of human lung donors are composed of basal or “progenitor cells” and are grown as nondifferentiated monolayers, which renders them a less ideal model. Farther removed from HBEC is the Ad12/SV40-immortalized BEAS-2B cell line, which was originally cloned from HBEC (Lechner et al., 1982). To corroborate the data obtained using HBEC and BEAS-2B, experiments were performed using HBEC cultured for 5 weeks at ALI, which resembles the structure of the airway epithelium in vivo. At this time, the epithelium was highly differentiated and contained specialized elements including ciliated and goblet cells. Consistent with HBEC grown in submersion culture and the BEAS-2B cell line, ERK1 and ERK2 were also phosphorylated in unstimulated ALI cultures (Fig. 2G). Moreover, exposure of these cells to formoterol (1 nM; 60 minutes) and EGF (10 ng/ml; 60 minutes) decreased and increased, respectively, the phosphorylation of both ERK isoforms (Fig. 2G).
Formoterol Promoted ERK1/2 Dephosphorylation in Airway Epithelial Cells Determined by ECL
The ability of formoterol and EGF to modulate ERK1/2 phosphorylation was also investigated by ECL using the MesoScale Discovery platform. This alternative, quantitative approach, which measures the fraction of the total ERK1/2 pool in an active state, confirmed the results obtained by Western blotting. Thus, in BEAS-2B cells, formoterol dephosphorylated ERK1/2 in a time-dependent manner (t1/2 = 6.7 minutes) that by 30 minutes had plateaued to a new steady state that was at ∼25% of the unstimulated level (Fig. 2H). Similarly, formoterol reduced the pERK1/2 level in HBEC to 20% of the unstimulated level at 60 minutes, whereas in HEK293β2 cells, it was increased by 3.3-fold at 5 minutes (Fig. 2I). Epidermal growth factor (10 ng/ml; 30 minutes) increased ERK1/2 phosphorylation in all three cell models, which was again consistent with the immunoblotting results (Fig. 2I). Of note, ECL established that basal ERK1/2 phosphorylation in HEK293β2 cells was 25- and 40-fold lower than in BEAS-2B cells and HBEC, respectively (Fig. 2I). Nevertheless, formoterol (1 nM) still enhanced ERK1/2 phosphorylation in HEK293β2 cells pretreated (10 minutes) with a concentration of EGF (1 ng/ml) that increased pERK1/2 to levels similar to those measured in unstimulated BEAS-2B cells (Fig. 2J).
β2-Adrenoceptor Agonists Promoted ERK1/2 Dephosphorylation in Airway Epithelial Cells by Canonical, cAMP-dependent Signaling
In HBEC, the adenylyl cyclase activator, forskolin (10 μM), promoted a time-dependent and sustained dephosphorylation of ERK1/2 (t1/2 ∼11 minutes) that had declined and plateaued to 25% of the unstimulated level by 30 minutes (Fig. 3A). A cAMP-dependent mechanism was also implied in BEAS-2B cells treated with the selective E-prostanoid (EP)2- and EP4-receptor agonists ONO-AE1-259 and ONO-AE1-329, respectively (Suzawa et al., 2000). These ligands promote canonical, cAMP-dependent signaling in human airway epithelial cells but are partial agonists (Joshi et al., 2021) and were, therefore, used in cells pretreated with the PDE4 inhibitor, RNO (1 μM). Exposure of BEAS-2B cells to ONO-AE1-259 and ONO-AE1-329 (both 1 μM; 60 minutes) reduced basal pERK1/2 levels, which was blocked by the EP2- and EP4-receptor antagonists TG4-155 (Jiang et al., 2012) and L-161,982 (Machwate et al., 2001), respectively (both 1 μM; 30 minutes pretreatment) (Fig. 3B).
The role of canonical cAMP signaling was further interrogated by infecting HBEC and BEAS-2B cells with adenovirus vectors that encode PKIα, a selective inhibitor of PKA (Glass et al., 1986), or a control protein, GFP, at MOIs ranging from 1 to 100 (Fig. 3, C and D). At an MOI of 30, >95% of cells expressed either the GFP or PKIα transgenes (Meja et al., 2004) (Fig. 3, C and D). Moreover, the activation by formoterol (1 nM; 6 hours) of a CRE reporter stably transfected into BEAS-2B cells was prevented by PKIα with MOIs causing 50% and 95% inhibition of 2.6 and 10, respectively, whereas GFP-expressing cells were unaffected. In contrast, in the presence of RNO (1 µM), formoterol-induced (1 nM) and forskolin-induced (10 μM) cAMP accumulation in BEAS-2B cells was augmented by PKIα (Fig. 3D). Consistent with a previous report, this effect may be due to the inhibition of a PKA-activated PDE that is distinct from PDE4 (Violin et al., 2008). In both epithelial cell models, the expression of PKIα but not GFP (MOI = 25) abolished formoterol-induced ERK1/2 dephosphorylation (Fig. 3, E and F). Comparable results were obtained in HBEC and BEAS-2B cells treated with salmeterol (Supplemental Fig. 5), indicating that LABA-induced ERK1/2 dephosphorylation in human airway epithelial cells was dependent upon canonical, cAMP-dependent signaling and involved the activation of PKA.
Induction of DUSP1 Contributed to LABA-Induced ERK1/2 Dephosphorylation in Airway Epithelial Cells
In total, 10 MAP kinase phosphatases (MKPs) have been identified that form part of the larger DUSP superfamily (Theodosiou and Ashworth, 2002). As ERK1 and ERK2 are substrates for several MKPs, it was hypothesized that one or more of these enzymes may contribute to β2-adrenoceptor–mediated ERK1/2 dephosphorylation in airway epithelial cells. RNA-sequencing identified 18 DUSP mRNAs that were common to BEAS-2B cells and HBEC (Fig. 4A). Of those, only DUSP1 (MKP1), DUSP4 (MKP2), DUSP8 (hVH5), and DUSP10 (MKP5) were induced by LABAs in both cell models (Fig. 4A). Since DUSP1 mRNA was profoundly increased by formoterol at 1 hour (Fig. 4A) and is an established negative regulator of MAP kinase signaling (Sun et al., 1993; Lang and Raffi, 2019), its role in mediating ERK1/2 dephosphorylation was explored.
The induction of DUSP1 determined by gene expression profiling was validated by PCR over a time frame of 0–6 hours. Formoterol (1 nM) rapidly increased DUSP1 expression in BEAS-2B cells. This was detectable at 15 minutes (a time when dephosphorylation of ERK1/2 was evident), peaked at 60 minutes, and had returned to unstimulated levels by 4–6 hours (Fig. 4B). In BEAS-2B and HBEC, DUSP1 protein levels increased with kinetics that mirrored the changes in mRNA; for salmeterol (100 nM), the level increased rapidly, peaked at 60 minutes, and had returned to baseline by ∼4 hours (Fig. 4, C and D). At the 1-hour time point, formoterol and salmeterol increased DUSP1 expression in BEAS-2B cells in a concentration-dependent manner, with EC50 values of 78 pM and 2.5 nM, respectively (Fig. 4, E and F). The induction of DUSP1 mRNA by salmeterol and formoterol in both epithelial cell models was mimicked by forskolin (10 µM). Moreover, the effects of formoterol (HBEC) and salmeterol (BEAS-2B cells) were prevented after infection with Ad5.CMV.PKIα (MOI = 25) but not the control virus (Fig. 4G). Pretreatment (30 minutes) of BEAS-2B cells with ICI 118,551 (1 μM) inhibited the increase of DUSP1 induced by salmeterol, whereas the effect of forskolin was unchanged (Fig. 4G).
Transfection of BEAS-2B cells with DUSP1-targeting siRNAs (25 nM) markedly reduced (by >95%) the peak expression of DUSP1 protein induced by salmeterol (100 nM) that occurred at 1 hour and, in the same cells, reduced ERK1/2 dephosphorylation by 30%–50%; in contrast, a control siRNA (25 nM) had no significant effect (Fig. 4H). Collectively, these results indicated that the PKA-dependent induction of DUSP1 played a minor role in β2-adrenoceptor–mediated ERK1/2 dephosphorylation in human airway epithelial cells.
Phosphorylation of C-Raf Contributed to LABA-Induced ERK1/2 Dephosphorylation in BEAS-2B Cells
The protein kinases B-Raf and C-Raf activate and inhibit, respectively, downstream MEK1/2-ERK1/2 signaling in response to an increase in cAMP (Dumaz and Marais, 2005). Given the modest role of DUSP1 in LABA-induced ERK1/2 dephosphorylation (vide supra), the possible contribution of C-Raf in this response was explored using BEAS-2B cells. There are 13 known phosphorylation sites in C-Raf that regulate catalysis (Fig. 5A). Three of those residues (Ser43, Ser233, and Ser259) independently inhibit kinase activity when phosphorylated by PKA by preventing C-Raf from interacting with the small GTP-binding protein, Ras (Cook and McCormick, 1993; Dumaz and Marais, 2005). To determine whether Ser43 and Ser259 were phosphorylated by formoterol with kinetics that would account for ERK1/2 dephosphorylation, Western blotting was performed using phosphospecific antibodies. The phosphorylation of Ser233 was not determined because antibodies are unavailable. The amino acids Ser338 and Ser621 were also probed, as they are markers of C-Raf activation (Yip-Schneider et al., 2000; Ghosh et al., 2015; Takahashi et al., 2017a).
In BEAS-2B cells, Western blotting determined Ser259, Ser338, and Ser621 were phosphorylated constitutively in unstimulated cells relative to Ser43 (Fig. 5B). Formoterol (1 nM) rapidly increased the phosphorylation of Ser43 and Ser259, with similar t1/2 values (∼30 seconds). These events preceded the dephosphorylation of MEK1/2 (t1/2 = 1.2 minutes) and ERK1/2 (t1/2 = 6 minutes), which was consistent with the sequential activation of the C-Raf-MEK1/2-ERK1/2 signaling cascade (Fig. 5, B and C). Conversely, formoterol reduced the basal phosphorylation of the C-Raf activation marker Ser338 with slower kinetics (t1/2 = 4.7 minutes), whereas the level of pSer621 was unchanged (Fig. 5, B and C). A different profile of serine phosphorylation changes was seen in BEAS-2B cells treated with EGF (10 ng/ml; 30 minutes). As predicted, this mitogen produced a robust activation of the C-Raf-MEK1/2-ERK1/2 signaling cascade, which was associated with increased levels of pSer43 and pSer338 and reduced phosphorylation of Ser259 (Supplemental Fig. 6).
At 60 minutes, changes in formoterol-induced serine phosphorylation were prevented in cells infected with Ad5.CMV.PKIα but not the GFP-expressing control virus (Fig. 5, D and E). The dephosphorylation of MEK1/2 and ERK1/2 was also blocked by this intervention, indicating that formoterol had inactivated C-Raf in a PKA-dependent manner and attenuated downstream MAP kinase signaling.
In contrast to formoterol, the MEK1/2 inhibitors, U0126 and PD 098059 (both 10 μM), had no effect on the phosphorylation status of C-Raf or MEK1/2 but markedly reduced basal ERK1/2 phosphorylation. PD 098059, which selectively targets MEK1 (Alessi et al., 1995; Dudley et al., 1995), was less effective than U0126 (Fig. 5D), suggesting that MEK2 was the dominant isoform that controlled ERK1/2 activity in BEAS-2B cells.
Formoterol Inhibited the Expression of EGR1 in Airway Epithelial Cells
Extracellular signal-regulated kinases 1 and 2 are the terminal kinases in the Ras-C-Raf-MEK1/2 signaling cascade and regulate a diverse repertoire of downstream targets, including the transcription factor Elk-1 (Ünal et al., 2017). To establish a functional correlate of LABA-induced ERK-1/2 dephosphorylation, the activation of Elk-1 and the expression of an Elk-1–sensitive gene, EGR1 (Khachigian and Collins, 1998; Pagel and Deindl, 2011), which is known to be regulated by β2-adrenoceptor agonists in airway epithelial cells (Yan et al., 2018), were determined. EGR1 is a master regulator of cell cycle transition (Meloche and Pouysségur, 2007); is highly expressed in the lungs of individuals with a variety of respiratory diseases (Zhang et al., 2000; Ning et al., 2004; Yasuoka et al., 2009), including asthma (Goleva et al., 2008); and plays a central role in fibrosis and wound healing (Wu et al., 2009; Bhattacharyya et al., 2013).
In unstimulated BEAS-2B cells, HBEC, and ALI cultures EGR1 was constitutively expressed, in agreement with previous reports (Yan et al., 2018; Joshi et al., 2019). Formoterol (1 nM) reduced basal EGR1 mRNA and protein expression in all three cell models by >50% after 1 hour and/or 2 hours of exposure (Fig. 6, A and B). Kinetic analysis in BEAS-2B cells determined that repression of EGR1 transcripts was time-dependent (t1/2 ∼50 minutes) and had declined to ∼50% and ∼13% of the unstimulated level by 1 hour and 2 hours, respectively (Fig. 6C). The dual MEK1/2 inhibitor U0126 (10 µM), alone and in combination with formoterol, likewise repressed EGR1 with kinetics that were similar to formoterol (Fig. 6C). Baseline EGR1 mRNA was also reduced (by ∼50% at 1 hour) in HEK293β2 cells exposed to U0126 (Fig. 6D), indicating that this gene was regulated similarly across these cell types. In contrast, formoterol (1 nM) upregulated EGR1 expression in HEK293β2 cells at 1 hour (Fig. 6D), which was consistent with its ability to enhance, rather than inhibit, basal ERK1/2 phosphorylation (Fig. 1).
At 2 hours, U0126 and PD 098059 (both 10 μM), had markedly reduced pElk-1 levels and the constitutive expression of EGR1 (Fig. 6E), although PD 098059 was a relatively weak inhibitor, as it was of ERK1/2 phosphorylation (Figs. 5 and 6E). Formoterol (1 nM) also dephosphorylated ERK1/2 and Elk-1 and repressed EGR1 protein expression. These effects were abolished in cells infected with Ad5.CMV.PKIα but not the GFP-expressing control virus, establishing a role for canonical cAMP signaling in β2-adrenoceptor–mediated EGR1 repression (Fig. 6E). In contrast, and consistent with the data in Fig. 4H, siRNA-induced DUSP1 knockdown had either a modest or no effect in protecting cells against these formoterol-induced changes that depended on the outcome (i.e., ERK1/2 vs. Elk-1 and EGR1) measured (Fig. 6E).
Distinct Effects of Carvedilol, Carazolol, and Alprenolol on ERK1/2 Phosphorylation between HEK293β2 and BEAS-2B Cells
Carvedilol is a biased agonist at the β2-adrenoceptor and apparently promotes ERK1/2 phosphorylation in HEK239β2 cells by noncanonical signaling that involves the recruitment of βArr2 (Wisler et al., 2007). The data presented in Fig. 7A are consistent with those findings. Thus, carvedilol (10 µM) produced a rapid and transient increase in ERK1/2 phosphorylation that peaked at 5 minutes (∼4.5-fold increase), declined, reaching a nadir at ∼15 minutes, and then increased again to a new steady state that at 60 minutes was 3.5-fold above the unstimulated levels (Fig. 7A). Two structurally related β2-adrenoceptor “antagonists,” carazolol and alprenolol (both 10 μM), also increased ERK1/2 phosphorylation in HEK293β2 cells (responses peaked between 1 and 5 minutes), but their effects were transient and had returned to baseline levels by 15 minutes (Fig. 7, B and C). Pretreatment (30 minutes) of HEK293β2 cells with ICI 118,551 (100 nM) abolished carvedilol-, alprenolol-, and carazolol-induced ERK1/2 phosphorylation (Supplemental Fig. 7).
Unlike in HEK293β2 cells, basal ERK1/2 phosphorylation in BEAS-2B cells was unaffected by carvedilol, carazolol, or alprenolol (each 10 μM) after 1, 5, 15, 30, and 60 minutes of treatment (Fig. 7, D–F).
βArr2 Constrained Formoterol-Induced cAMP Accumulation and ERK1/2 Dephosphorylation in BEAS-2B Cells
The inability of carvedilol and related ligands to promote ERK1/2 phosphorylation in BEAS-2B cells was enigmatic, given that epithelial cells express a high density of efficiently coupled β2-adenoceptors (Davis et al., 1990; Penn et al., 1994; Kelsen et al., 1995) at which partial agonists are predicted to display efficacy (Yan et al., 2018). Moreover, ARRB2 transcripts determined by RNA-sequencing were abundant in BEAS-2B cells and HBEC (TPMs: ∼35 and ∼9, respectively) and predominated (≥30-fold) over ARRB1, the other nonvisual arrestin (Fig. 8A). In BEAS-2B cells, these findings were validated at the protein level, with βArr2 being identified by Western blotting (Fig. 8B). A band corresponding to βArr1 was also labeled in BEAS-2B cells. However, this was only detected using an ultrasensitive chemiluminscent substrate (i.e., SignalFire Elite), which can detect proteins in the femtogram range (Supplemental Fig. 8A). In contrast, a very strong βArr1 signal was detected in HEK293β2 cells (Supplemental Fig. 8B). In both cell types, gene silencing confirmed that these bands were βArr1 (Supplemental Fig. 8).
To examine whether βArr2 played a role in β2-adrenoceptor–mediated signaling in BEAS-2B cells, three clones (C1, C2, and C3) deficient in this protein were derived by using CRISPR/Cas9 genome editing technology (Fig. 8B). In parental and clonal cells pretreated with the PDE4 inhibitor RNO (1 μM; 30 minutes), formoterol increased cAMP mass in a concentration-dependent manner with similar EC50 values (2.2 nM and 2.8 nM, respectively) (Fig. 8C). However, the change in maximal response was 2-fold greater in cells deficient in βArr2 (Fig. 8C). This effect was replicated when the time course of cAMP accumulation was determined in response to a supramaximally effective concentration of formoterol (100 nM), although the kinetics were similar (t1/2 = 1.5–2 minutes) in both clonal and parental cells (Fig. 8C). In contrast, ARRB2 deletion had no effect on RNO- or forskolin-induced cAMP generation (Fig. 8E), confirming the selectivity of βArr2 in promoting agonist-induced β2-adrenoceptor desensitization (Fig. 8D).
In unstimulated ARRB2−/− cells, ERK1 and ERK2 remained partially phosphorylated, although Western blotting and ECL determined that the level was significantly (P < 0.05) lower (∼30%) than in wild-type cells (Fig. 8F). Nevertheless, in all clones, formoterol (1 nM) produced a time-dependent and sustained dephosphorylation of ERK1/2 that had plateaued at 60 minutes to a level that was 25% of the control (Fig. 8G; Supplemental Fig. 9). Kinetically, the rate of dephosphorylation occurred more rapidly in ARRB2-deficient cells (t1/2 = 3.7 minutes) than in their parental counterparts (t1/2 = 11.0 minutes) (Fig. 8G), presumably because βArr2-mediated desensitization had been compromised (Fig. 8, C and D). Western blotting and ECL determined that the deletion of ARRB2 had no effect on ERK1/2 phosphorylation induced by EGF (10 ng/ml; 30 minutes) (Fig. 8, F and G).
Effect of Formoterol on the Expression of Signaling Components Required for βArr- and Gβγ-Dependent ERK1/2 Activation
The possibility that key signaling elements were either absent or expressed at limiting levels in untreated human primary airway epithelial cells such that a functional βArr2-signaling complex could not assemble was examined by RNA-sequencing. This analysis was extended to include components of the Gβγ-signaling pathway, which can promote ERK1/2 phosphorylation in a βArr-independent manner (Crespo et al., 1995; O'Hayre et al., 2017).
As shown in Supplemental Fig. 10, mRNAs encoding 27 signaling elements implicated in β2-adrenoceptor–mediated ERK1/2 phosphorylation were identified in untreated HBEC. These included βArr1, βArr2, Ras isoforms, C-Raf, MEK1, MEK2, ERK1, and ERK2, which are involved in βArr-signaling, and 16 components of the Gβγ-signaling pathway (i.e., Gβ1-β5, Gγ3-γ5, Gγ10-γ12, the proto-oncogene tyrosine-protein kinase, Src, the Src homology 2 domain-containing) transforming proteins, Shc1 and Shc2, and the son-of-sevenless guanine nucleotide exchange factors, Sos1 and Sos2). A majority (22) of these were expressed at levels similar to or greater than transcripts encoding the β2-adrenoceptor (Supplemental Fig. 10A).
Treatment of HBEC with formoterol (1 nM) for 1, 2, 6, and 18 hours did not induce or repress any of these mRNAs relative to vehicle-treated cells matched for time (Supplemental Fig. 10B). Likewise, formoterol failed to induce mRNAs encoding Gγ subunits that were not detected in unstimulated HBEC (i.e., Gγ1, Gγ2, Gγ6-8) under identical experimental conditions. Comparable data were derived from BEAS-2B cells treated with formoterol (data not shown; see GSE115830).
Histamine and Carbachol Promoted Gq-Dependent ERK1/2 Phosphorylation in BEAS-2B Cells
Additional studies were performed to determine whether ERK1/2 could be phosphorylated in airway epithelial cells by activators of GPCRs that signal independently of Gs. For this purpose, the airway-relevant agonists histamine and carbachol were examined together with the signaling roles of Gq and βArr2. Exposure of BEAS-2B cells to histamine or carbachol (both 10 μM) produced a transient increase in ERK1/2 phosphorylation (Fig. 8, H and I). The responses induced by both agonists peaked at 5 minutes (2- to 2.5-fold increase relative to vehicle matched for time) and then gradually declined over the next 25 minutes toward the baseline levels (Fig. 8, H and I). ERK1/2 phosphorylation was abolished in BEAS-2B cells pretreated (30 minutes) with mepyramine or atropine (both 1 μM), indicating that histamine and carbachol were acting through the H1-receptor and a muscarinic receptor subtype, respectively (Fig. 8, H and I). Likewise, histamine- and carbachol-induced ERK1/2 phosphorylation was completely inhibited in cells pretreated (60 minutes) with the Gq inhibitor YM-254890 (1 μM) (Fig. 8, J and K), whereas the effect of EGF (10 ng/ml; 30 minutes) was unchanged (Supplemental Fig. 11). Mepyramine, atropine, and YM-254890 did not affect basal pERK1/2 levels (Supplemental Fig. 11).
In ARRB2−/− BEAS-2B cells, the time to peak and magnitude of histamine- and carbachol-induced ERK1/2 phosphorylation was unchanged relative to parental cells (Fig. 8, L and M). However, the rate of ERK1/2 dephosphorylation was protracted. For histamine, the t1/2 was increased from 6.1 to 12.9 minutes in parental and clonal cells, respectively (Fig. 8L). A similar, but less pronounced, increase in t1/2 was seen with carbachol (parental: 4.5 minutes; clonal: 7.9 minutes) (Fig. 8M), implying that βArr2 had exerted a negative regulatory influence on histamine H1- and muscarinic receptor–mediated signaling.
Discussion
Chronic β2-adrenoceptor agonist monotherapy in asthma is associated with adverse clinical outcomes that may be partially driven by changes in proinflammatory gene expression. It has been proposed that this occurs principally in airway epithelial cells and involves the βArr2-dependent nucleation of the C-Raf-MEK1/2-ERK1/2 signaling complex (Nguyen et al., 2017). However, in the present study, the LABAs formoterol and salmeterol paradoxically reduced pERK1/2 and pMEK1/2 levels in airway epithelial cells by activating the Gαs-adenylyl cyclase-cAMP signaling cascade (Fig. 9). Mechanistically, evidence for two PKA-regulated mechanisms was obtained: disruption of Ras-C-Raf complex assembly and induction of DUSP1. At a genomic level, ERK1/2 dephosphorylation leads to the inactivation of the transcription factor Elk-1 and repression of a representative Elk-1–regulated gene, EGR1. Together, these findings indicate that acute exposure of airway epithelial cells to LABAs does not promote βArr2-dependent ERK1/2 phosphorylation and downstream gene expression changes; instead, cAMP-regulated mechanisms predominate, consistent with the conventional view (Zhang et al., 2005; Yan et al., 2018).
Antithetical Regulation of ERK1/2 Activity by β2-Adrenoceptor Agonists
In three human airway epithelial cell models, the structurally dissimilar LABAs formoterol, salmeterol, and indacaterol promoted a time-dependent dephosphorylation of ERK1/2 that was mediated by the β2-adrenoceptor. Based on these data, the likelihood that ERK1/2 would be similarly inactivated in the airway epithelium of asthmatic subjects after taking an inhaled β2-adrenoceptor agonist should be entertained. In contrast, and consistent with several previous reports (Shenoy et al., 2006; van der Westhuizen et al., 2014; Luttrell et al., 2018), formoterol and salmeterol increased basal ERK1/2 phosphorylation in HEK293β2 cells. Thus, the impact of β2-adrenoceptor agonists on MAP kinase signaling is cell type–dependent and should not be generalized. Indeed, the β2-adrenoceptor–mediated regulation of ERK1/2 activity is not even consistent across human airway epithelial cell lines (Nishimura et al., 2002), indicating the need to exercise caution in selecting a suitable system to model human primary cells.
Mechanistically, formoterol- and salmeterol-induced ERK1/2 dephosphorylation in BEAS-2B cells and HBEC was abolished by PKIα, suggesting that this was mediated by canonical, cAMP-dependent signaling. The adenylyl cyclase activator forskolin and the EP2- and EP4-receptor agonists ONO-AE1-259 and ONO-AE1-329, respectively, likewise decreased pERK1/2 levels, which strengthens this conclusion. EP4-receptor agonists acting via the cAMP-PKA axis also decrease ERK1/2 activity in chondrocytes and neutrophils (Fushimi et al., 2007; Mizuno et al., 2014), which is consistent with the signaling paradigm reported here. However cAMP-independent ERK1/2 phosphorylation and dephosphorylation have also been reported (Gerits et al., 2008), emphasizing again the context dependence of this response.
Early growth response 1 is one of several genes regulated by Elk-1 and related E twenty-six domain transcription factors. Many of these are downstream substrates of ERK1/2 (Buchwalter et al., 2004; Ünal et al., 2017), and their expression is regulated by interventions that change ERK1/2 activity. In HEK293β2 cells and all epithelial cell models, MEK1/2 inhibitors reduced constitutive Elk-1 phosphorylation and repressed EGR1 gene expression, confirming that basal Elk-1-EGR1 signaling was maintained by ERK1/2. In contrast, formoterol increased and decreased Elk-1–regulated EGR1 expression in HEK293β2 and airway epithelial cells, respectively, consistent with its opposing effects on ERK1/2 phosphorylation.
The role of EGR1 in asthma pathology and the consequences of repression produced by formoterol are unclear. A beneficial effect might be predicted given its association with several human fibrotic pulmonary disorders and that airway inflammation and mucus hypersecretion after exposure to particulate matter were inhibited in EGR1 knockout mice (Xu et al., 2018). However, a reduction in EGR1 could be detrimental under certain circumstances. Indeed, the lung pathology that develops in transgenic mice that overexpress transforming growth factor β was more severe in animals that also lacked EGR1 (Kramer et al., 2009). Given the central role of ERK1/2 in cell cycle regulation (Meloche and Pouysségur, 2007), one effect of chronic β2-adrenoceptor agonist therapy could be to arrest cell growth and repair by inhibiting constitutive gene expression programs that are actively maintained by pERK1/2.
Multiple Mechanisms Contribute to β2-Adrenoceptor–Mediated ERK1/2 Dephosphorylation Human Airway Epithelial Cells
BEAS-2B cells were used as a representative epithelial cell model to explore how formoterol and related LABAs caused ERK1/2 dephosphorylation. Initially, the participation of DUSP1 was investigated because it is upregulated by β2-adrenoceptor agonists (Manetsch et al., 2012; Tsvetanova and von Zastrow, 2014; Kang et al., 2016), is a classic target for the PKA-regulated transcription factor CRE-binding protein (Zhang et al., 2005), and can dephosphorylate ERK1/2 in intact cells (Caunt and Keyse, 2013; Moosavi et al., 2017). DUSP1 was also rapidly induced in BEAS-2B cells and primary HBEC over the time frame when ERK1/2 dephosphorylation was ongoing. Gene silencing established that, under conditions of >95% DUSP1 protein knockdown, β2-adrenoceptor–mediated ERK1/2 dephosphorylation was inhibited by ∼30%–50%, although downstream Elk-1 phosphorylation and EGR1 repression were unaffected. Although the ∼5% DUSP1 remaining in these cells may have been sufficient to ensure that most of the ERK1/2 pool was still dephosphorylated in response to agonist, other mechanisms were implicated. The observation that basal MEK1/2 phosphorylation was reduced by formoterol in a PKIα-sensitive manner indicates that a signaling component(s) upstream of ERK1/2 was also a PKA substrate. Indeed, the kinetics of MEK1/2 dephosphorylation were rapid (t1/2 = 1.2 minutes) in BEAS-2B cells and consistent with the slower and causal inactivation of ERK1/2 (t1/2 = 6.7 minutes). There is good evidence that C-Raf regulates ERK1/2 activity by physically linking Ras to MEK1/2 (Dougherty et al., 2005; Lavoie and Therrien, 2015). Moreover, PKA can disrupt this nucleation by phosphorylating C-Raf at Ser43 and Ser233/Ser259, which weakens the Ras-C-Raf interaction by steric hindrance and stabilizes the kinase in an inactive conformation, respectively (Dhillon et al., 2002; Dumaz and Marais, 2003; Dougherty et al., 2005; Lavoie and Therrien, 2015). In agreement with these prior data, formoterol similarly enhanced the basal phosphorylation of C-Raf in BEAS-2B cells at Ser43 and Ser259. This occurred rapidly (t1/2 ∼0.5 minutes), was dependent on PKA, and preceded the dephosphorylation of MEK1/2 and ERK1/2. Formoterol also reduced the basal phosphorylation of a C-Raf activation marker, Ser338 (Takahashi et al., 2017a). Taken together, these data suggest that the primary means by which β2-adrenoceptor agonists inactivated MEK1/2-ERK1/2-Elk-1-EGR1 signaling in airway epithelial cells was by disrupting the link between Ras and MEK1/2 through the PKA-dependent phosphorylation of C-Raf (Fig. 9).
Agonist-Induced Desensitization of the β2-Adrenoceptor on Human Airway Epithelial Cells Was Mediated by βArr2
In BEAS-2B cells and HBEC, βArr2 was the predominant nonvisual arrestin, with a relatively minor representation from βArr1. Functionally, formoterol-induced cAMP accumulation was enhanced in BEAS-2B cells deficient in βArr2 when compared with their parental counterparts. These data support the established role of βArr2 in mediating agonist-induced homologous desensitization of the β2-adrenoceptor, although it remains unclear whether βArr1 is more or less important in this regard. Although both arrestins can restrain coupling of the agonist-bound β2-adrenoceptor to Gs, the assembly of the C-Raf-MEK1/2-ERK1/2 signaling complex is suggested to be more dependent on βArr2 (Penn et al., 2001; Tohgo et al., 2003; Luttrell et al., 2018). Nevertheless, βArr2 has been reported to promote β2-adrenoceptor desensitization in murine embryonic fibroblasts (Baillie et al., 2007), consistent with the results of the present study.
No Evidence for β2-Adrenoceptor–Mediated, βArr2-Dependent ERK1/2 Phosphorylation in Human Airway Epithelial Cells
Having established that βArr2 was expressed and functional in BEAS-2B cells, the possibility that ERK1/2 is activated by a noncanonical, G-protein–independent mechanism was investigated. Carvedilol, a partial β2-adrenoceptor agonist biased toward βArr2 (Wisler et al., 2007; Liu et al., 2012), was used for this purpose, as it activates ERK1/2 in HEK293β2 cells (Wisler et al., 2007; Luttrell et al., 2018; this study). However, consistent with data gained in another epithelial cell line (Peitzman et al., 2015), carvedilol did not affect basal pERK1/2 levels in BEAS-2B cells. Although this may indicate the absence of βArr2-dependent signaling, relatively low β2-adrenoceptor density (cf. HEK293β2 cells) may have rendered carvedilol inactive. Indeed, the law of mass action dictates that a decrease in receptor number will reduce potency and efficacy, especially of a partial agonist like carvedilol. This explanation suggests that ERK1/2 phosphorylation may be realized with biased agonists of high intrinsic efficacy and, possibly, with carvedilol under conditions of β2-adrenoceptor overexpression. Low receptor density could also explain why alprenolol (Wisler et al., 2007; Liu et al., 2012) and carazolol (unpublished) promoted Gαs-dependent ERK1/2 phosphorylation in HEK293β2 cells but not in the BEAS-2B cell line.
The increase in formoterol-induced cAMP accumulation in βArr2-deficient cells was associated with an accelerated rate at which ERK1/2 was dephosphorylated (t1/2 was reduced from 9.7 to 4 minutes). Although this kinetic discrepancy is consistent with compromised β2-adrenoceptor desensitization, such data can also be explained by the ablation of βArr2-dependent ERK1/2 phosphorylation that occurs rapidly (≤5 minutes) in some cell types after agonist stimulation (Shenoy et al., 2006). However, in BEAS-2B cells and HBEC, this interpretation is inconsistent with the finding that PKIα abolished ERK1/2 dephosphorylation without revealing latent activation at any time point.
Additional signaling roles for βArr2 were explored by investigating the effect of carbachol and histamine on ERK1/2 phosphorylation in ARRB2−/− and wild-type BEAS-2B cells. Histamine and carbachol produced transient increases in ERK1/2 phosphorylation that were dependent upon Gq. In agreement with muscarinic M3-receptor signaling in HEK293 cells (Luo et al., 2008), neither carbachol- nor histamine-induced ERK1/2 phosphorylation was inhibited in ARRB2−/− BEAS-2B cells. In fact, in both experiments, the rate of decline of the pERK1/2 signal was reduced in cells lacking βArr2. This extended kinetic is consistent with exaggerated Gq signaling, presumably because the histamine H1- and muscarinic receptors were no longer susceptible to homologous desensitization. These data are contrary to the siRNA-mediated knockdown of βArr2 reported in myometrial cells, in which histamine H1-receptor–mediated ERK1/2 activation was abolished (Brighton et al., 2011), indicating that this form of GPCR regulation is system-dependent.
Cell Type– and Context-Dependent Regulation of ERK1/2 Activity by β2-Adrenoceptor Agonists
Why are LABAs unable to promote βArr2-dependent ERK1/2 phosphorylation in airway epithelial cells? The idea that essential components required to nucleate a functional βArr-signaling complex were absent or limiting was considered but seems unlikely. Indeed, mRNAs encoding 27 key proteins implicated in βArr- and Gβγ-regulated signaling were expressed in HBEC and BEAS-2B cells at levels similar to, or in excess of, ADRB2, which encodes the β2-adrenoceptor. Moreover, EGF promoted robust ERK1/2 phosphorylation in all three epithelial cell models, in which many of these same signaling intermediates are obligatory. Alternatively, basal phosphorylation of ERK1/2 in airway epithelial cells may have been maximal. Indeed, ECL determined that pERK1/2 levels were considerably higher (25- to 40-fold) in BEAS-2B cells and HBEC than in HEK293β2 cells. However, that idea could not be reconciled with the activation of ERK1/2 by EGF, TNFα, PMA, histamine, and carbachol. The results with the latter two agonists are noteworthy because they illustrate that the G-protein to which a given GPCR preferentially couples dictates whether ERK1/2 activity is increased or decreased. Furthermore, formoterol augmented ERK1/2 phosphorylation in HEK293β2 cells, in which the baseline was elevated by EGF to a level similar to that measured in unstimulated BEAS-2B cells. The reason for high, constitutive C-Raf/MEK1/2/ERK1/2 signaling in unstimulated BEAS-2B cells (cf. HEK293β2s) is unknown. Cell culture conditions were similar, which suggests that this is not the cause. However, the release of an autocrine factor(s) that activates this pathway should be considered. Alternatively, sufficient βArr2 may have been plasma membrane–associated in unstimulated airway epithelia to assemble an active C-Raf–signaling complex independently of an agonist-bound GPCR (Terrillon and Bouvier, 2004; Jafri et al., 2006). The ∼30% reduction in basal pERK1/2 levels in clonal cells lacking βArr2 is consistent with this idea. Thus, β2-adrenoceptor agonists may be unable to increase ERK1/2 phosphorylation in human airway epithelial cells because of constitutive, βArr-dependent signaling.
The discrepancy between the findings reported here and those derived from murine models of asthma merits discussion. Notwithstanding a species difference, context-dependent factors may play a role. Indeed, β2-adrenoceptor agonists were administered to mice chronically or repeatedly prior to allergen challenge (Lin et al., 2012; Thanawala et al., 2013), which could have resulted in an adaptation of the signaling that would typically follow acute agonist exposure. In particular, chronic administration could promote genomic effects in the airway epithelium that modify the regulation of βArr2. However, mRNAs encoding key components involved in βArr2- and Gβγ-dependent ERK1/2 phosphorylation were unchanged in HBEC and BEAS-2B cells exposed to formoterol for up to 18 hours. Although these negative data do not necessarily extrapolate to mice subjected to chronic agonist exposure, they nevertheless argue against such a mechanism in human primary airway epithelial cells. They also raise the prospect that β2-adrenoceptor agonists exacerbate the asthma-like pathology in murine models of asthma by canonical mechanisms that regulate βArr2 activity rather than biased agonism. Indeed, cAMP increases ERK1/2 activity in many cell types in which βArr2 may play an indispensable role.
Conclusions
The experiments described herein failed to unearth evidence for β2-adrenoceptor–mediated, βArr2-dependent signaling in airway epithelia; neither conventional (e.g., formoterol) nor biased (e.g., carvedilol) agonists increased pERK1/2 levels, whereas robust phosphorylation was evident in HEK293β2 cells. In fact, β2-adrenoceptor agonists paradoxically dephosphorylated ERK1/2 in all three human cell models examined, including highly differentiated primary cultures. This antithetical behavior of formoterol and related ligands is striking and underscores the cell type dependence of β2-adrenoceptor–mediated signaling. It also emphasizes the need in drug discovery to evaluate lead candidates in the therapeutic target of interest or in systems in which receptor density and coupling to downstream effectors are similar. The failure of carvedilol to phosphorylate ERK1/2 in BEAS-2B cells, in which receptor number for this ligand may have been limiting, highlights this dilemma and has been acknowledged previously (Wisler et al., 2007; Luttrell et al., 2018). Thus, βArr2 may assemble a functional C-Raf-MEK1/2-ERK1/2 signaling complex in native human airway epithelial cells, but only in response to biased β2-adrenoceptor agonists with high intrinsic efficacy.
Acknowledgments
We thank Dr. Michel Bouvier (University of Montreal, Canada) for providing HEK293β2 cells.
Authorship Contributions
Participated in research design: Hamed, Giembycz.
Conducted experiments: Hamed, Joshi, Michi, Kooi.
Performed data analysis: Hamed, Joshi, Giembycz.
Wrote or contributed to the writing of the manuscript: Hamed, Giembycz, Michi.
Footnotes
- Received April 1, 2021.
- Accepted July 7, 2021.
This study was supported by the Canadian Institutes for Health Research [PJT 152904] and the Natural Sciences and Engineering Research Council [Discovery Grant RGPIN-2018-04312]. O.H. was a recipient of graduate scholarship awarded by The Lung Association of Alberta & Northwest Territories. Real-time PCR was facilitated by an equipment and infrastructure grant from the Canadian Fund of Innovation and the Alberta Science and Research Authority
The authors state no conflict of interest.
↵This article has supplemental material available at molpharm.aspetjournals.org.
Abbreviations
- ALI
- air-liquid interface
- Arr
- arrestin
- ARRB
- arrestin gene
- CRE
- cAMP response element
- DUSP
- dual-specificity phosphatase
- ECL
- electrochemiluminescence
- EGF
- epidermal growth factor
- EGR1
- early growth response 1
- ERK
- extracellular signal-regulated kinase
- GPCR
- G-protein–coupled receptor
- HBEC
- human bronchial epithelial cell
- ICI 118
- 551(2R3R)-1-[(7-methyl-23-dihydro-1H-inden-4-yl) oxy]-3-(propan-2-ylamino)butan-2-ol
- L-161
- 982N-[2-[4-[[3-butyl-5-oxo-1-[2-(trifluoromethyl)phenyl]-124-triazol-4-yl]methyl]phenyl]phenyl]sulphonyl-5-methyl-thiophene-2-carboxamide
- LABA
- long-acting β2-adrenoceptor agonist
- MAP
- mitogen-activated protein
- MEK
- mitogen-activated protein kinase kinase
- MKP
- mitogen-activated protein kinase phosphatase
- MOI
- multiplicity of infection
- ONO-AE1-259
- (Z)-7-[(1R2R3R5R)-5-chloro-3-hydroxy-2-[(E4S)-4-hydroxy-4-(1-prop-2-enylcyclobutyl)but-1-enyl]cyclopentyl] hept-5-enoic acid
- ONO-AE1-329
- 2-[3-[(1R2S3R)-3-hydroxy-2-[(E3S)-3-hydroxy-5-[2-(methoxymethyl)phenyl]pent-1-enyl]-5-oxo-cyclopentyl]sulphanyl propylsulphanyl] acetic acid
- p
- phospho
- PD 098059
- 2-(2-amino-3-methoxyphenyl) chromen-4-one
- PDE
- phosphodiesterase
- PKA
- cAMP-dependent protein kinase
- PKI
- cAMP-dependent protein kinase inhibitor
- PMA
- phorbol 12-myristate 13-acetate
- Raf
- rapidly accelerated fibrosarcoma
- Ras
- rat sarcoma
- RNO
- roflumilast N-oxide
- SFM
- serum-free medium
- t
- total
- t1/2
- half-life
- TG4-155
- (E)-N-[2-(2-methylindol-1-yl) ethyl]-3-(345-trimethoxyphenyl)prop-2-enamide
- TNFα
- tumor necrosis factor-α
- TPM
- transcripts per million
- U0126
- (2Z3Z)-23-bis[amino-(2-aminophenyl)sulphanylmethylidene] butane dinitrile
- YM-254890
- [(1R)-1-[(3S6S9S12S18R21S22R)-21-acetamido-18-benzyl-3-[(1R)-1-methoxyethyl]-4910121622-hexamethyl-15-methylidene-25811141720-heptaoxo-119-dioxa-47101316-pentazacyclodocos-6-yl]-2-methylpropyl](2S3R)-2-acetamido-3-hydroxy-4-methyl pentanoate
- Copyright © 2021 by The American Society for Pharmacology and Experimental Therapeutics