Specificity of arrestin subtypes in regulating airway smooth muscle G protein–coupled receptor signaling and function

Abstract

Arrestins have been shown to regulate numerous G protein–coupled receptors (GPCRs) in studies employing receptor/arrestin overexpression in artificial cell systems. Which arrestin isoforms regulate which GPCRs in primary cell types is poorly understood. We sought to determine the effect of β-arrestin-1 or β-arrestin-2 inhibition or gene ablation on signaling and function of multiple GPCRs endogenously expressed in airway smooth muscle (ASM). In vitro [second messenger (calcium, cAMP generation)], ex vivo (ASM tension generation in suspended airway), and in vivo (invasive airway resistance) analyses were performed on human ASM cells and murine airways/whole animal subject to β-arrestin-1 or -2 knockdown or knockout (KO). In both human and murine model systems, knockdown or KO of β-arrestin-2 relative to control missense small interfering RNA or wild-type mice selectively increased (40–60%) β₂-adrenoceptor signaling and function. β-arrestin-1 knockdown or KO had no effect on signaling and function of β₂-adrenoceptor or numerous procontractile GPCRs, but selectively inhibited M3 muscarinic acetylcholine receptor signaling (∼50%) and function (∼25% ex vivo, >50% in vivo) without affecting EC₅₀ values. Arrestin subtypes differentially regulate ASM GPCRs and β-arrestin-1 inhibition represents a novel approach to managing bronchospasm in obstructive lung diseases.—Pera, T., Hegde, A., Deshpande, D. A., Morgan, S. J., Tiegs, B. C., Theriot, B. S., Choi, Y. H., Walker, J. K. L., Penn, R. B. Specificity of arrestin subtypes in regulating airway smooth muscle G protein–coupled receptor signaling and function.

G protein–coupled receptors (GPCRs) are the principal transducers of pro- and anticontractile stimuli in airway smooth muscle (ASM) (1–3). GPCRs of the G_q class (i.e., GPCRs that activate the heterotrimeric G_q protein) promote ASM contraction primarily via activation of phospholipase C, which leads to an increase in intracellular calcium that promotes cross-bridge cycling. Alternatively, GPCRs that couple to G_s can antagonize ASM contraction via inhibition of G_q-induced calcium flux and by reducing cellular sensitivity to calcium. ASM contractile state is largely determined by the dynamic balance of signaling between G_q and G_s, and disruption of this balance is reflected in the airway disease asthma, in which exaggerated presentation of G_q-coupled receptor agonists (e.g., acetylcholine, cysteinyl leukotrienes, histamine) in the airway promotes bronchoconstriction. Most therapeutic strategies for bronchorelaxation restore the balance of signaling by either decreasing G_q signaling (muscarinic acetylcholine or cysteinyl leukotriene receptor antagonists), or augmenting G_s signaling (β-agonists).

Although β-agonists, acting on the β2-adrenoceptor (β₂AR), are the mainstay of bronchorelaxant therapy in airways diseases, their chronic use results in a decreased bronchoprotective effect and has been associated with loss of disease control in asthma (4–8), as reviewed by Penn et al. (3). Desensitization of the β₂AR on ASM cells may contribute to this phenomenon. Upon occupation of the β₂AR by β-agonists, the receptor is phosphorylated by G protein–coupled receptor kinases (GRKs) (9–11). This phosphorylation of GPCRs leads to reduced coupling to G proteins and promotes the binding of arrestins. Arrestins bound to the receptor cause steric hindrance of the receptor-G protein interaction, effectively uncoupling the receptor from G protein and promoting receptor internalization (12). Although these mechanisms of β₂AR desensitization and internalization have been intensely studied in a variety of artificial in vitro and cell-based systems, their applicability, or lack thereof, in more integrative, physiologic systems is poorly understood. Given that the competitive nature of ASM GPCRs dictates ASM contractile state and thus airway resistance, understanding how receptor responsiveness is (differentially) regulated in ASM may provide insight into new therapeutic strategies, for example targeting arrestins and GRKs (13, 14), that could either augment prorelaxant, or antagonize procontractile, signaling.

We have previously shown that overexpression of β-arrestin (βarr)-1 or βarr-2 causes β₂AR desensitization and attenuates β-agonist–induced signaling in primary human ASM cells (15). More recently, we have determined that βarr-2 knockout (KO) in mice augments β-agonist–induced signaling and function in ASM, using in vitro, ex vivo, and in vivo approaches (16). Similarly, knockdown of βarr-1/2 in primary human ASM cells also augmented β₂AR signaling. By contrast, responses to prostaglandin E2 (PGE₂) or cholinergic stimulation were unaffected by βarr-2 KO, βarr-1/2 knockdown (16) or GRK2/3 knockdown (13) indicating selectivity of arrestins for β₂AR-induced signaling and bronchodilation.

Although the sufficiency of βarr-2 inhibition to augment β₂AR ASM signaling and function has been established, the effect of specific βarr-1 inhibition/ablation on the responsiveness of β₂ARs, or other GPCRs, in ASM remains unknown. Resolving the differential regulation of ASM GPCRs by arrestin isoforms will help identify the optimal therapeutic strategy for antagonizing procontractile GPCRs while enhancing the function of prorelaxant GPCRs.

In the current study, we investigated the selectivity of βarr subtypes in the regulation of GPCR signaling in ASM. Our results demonstrate that the βarr-2 isoform preferentially desensitizes β₂AR signaling function in ASM, but does not regulate procontractile receptors in ASM. Conversely, βarr-1 does not desensitize the β₂AR but appears required for maintaining procontractile M3 muscarinic acetylcholine receptor (mAChR) function in ASM.

MATERIALS AND METHODS

Animal generation, care, and use

βarr-1^−/− and βarr-2^−/− mice were generated as described previously by Conner et al. (17) and Bohn et al. (18), respectively. All animal care and experimental protocols were reviewed and approved by the Institutional Animal Care and Use Committee at Duke University Medical Center and were carried out in accordance with the standards established by the U.S. Animal Welfare Acts. Male C57/BL6 mice aged 6–8 wk were purchased from The Jackson Laboratory (Bar Harbor, ME, USA). βarr-1- or βarr-2-KO male mice were bred in house. All mice were kept in a pathogen-free barrier facility.

Immunoblot analysis

Mouse lung or trachea was pulverized in liquid nitrogen and solubilized in RIPA buffer. Tissue lysates were buffered with Laemmli solution and sonicated, and equal amounts of total protein (as measured by a DC protein assay read at 750 nm; Bio-Rad, Hercules, CA, USA) were loaded and separated on 10% Tris-glycine polyacrylamide gels (Invitrogen, Carlsbad, CA, USA). Proteins were transferred to nitrocellulose membranes for immunoblotting. βarrs were detected with the polyclonal antibody A2CT (generous gift from Robert J. Lefkowitz, Duke University, Durham, NC, USA). A2CT preferentially recognizes βarr-2 over βarr-1 (19). Chemiluminescent detection was performed with horseradish peroxidase–coupled secondary antibody (Amersham Biosciences, Pittsburgh, PA, USA) and SuperSignal West Femto reagent (Pierce, Rockford, IL, USA). Chemiluminescence was quantified by a charge-coupled device camera (Syngene, Frederick, MD, USA); representative images are shown as inverted grayscale. For immunoblots of lysates from cultured human ASM cells, blots were probed with A1CT antibody (generous gift from Robert J. Lefkowitz) and an IRDye 680 secondary antibody (Li-Cor, Lincoln, NE, USA) conjugated with an infrared fluorophore. A1CT antibody preferentially recognizes βarr-1 over βarr-2 (19). Bands were visualized and signals (infrared emission) quantified directly using the Odyssey Infrared Imaging System (Li-Cor), as described previously (20).

Assessment of airway responsiveness: airway pressure time index and impedance

Airway responsiveness was measured in a terminal procedure using the forced oscillation technique or airway pressure time index (APTI). In brief, mice were anesthetized (sodium pentobarbital 85 mg/kg, i.p.), tracheotomized, paralyzed (pancuronium bromide 0.25 mg/kg, i.v.) and ventilated with 100% oxygen at 150 breaths per minute, constant volume of 8 ml/kg. For APTI, peak tracheal pressure was continuously acquired from a tracheal cannula side port. APTI was calculated as the sum of the post methacholine (MCh)-induced changes in peak tracheal pressure (relative to pre-MCh peak tracheal pressure) integrated with respect to time (30 s). APTI is a measure that others have validated for its ability to provide a reasonable index of airway responsiveness (21) as assessed by the more specific mechanical variables of resistance and compliance. For impedance measurements, a computer-controlled small animal ventilator (flexiVent; Scireq, Montreal, QC, Canada) was used that delivered a 3 s optimized pseudorandom signal containing frequencies ranging from 0.25 to 19.63 Hz, and the resultant total lung impedance signal contained information about the resistance and elastance properties of the lung from which Newtonian resistance was calculated using the constant phase model. A 3 cmH₂O positive end-expiratory pressure was used. Following ∼5 min of regular ventilation, a standard lung volume history was established followed by the acquisition of 3 baseline respiratory input impedance measurements. Bronchospasm was induced by jugular vein administration of increasing doses of MCh (50, 100, 200 μg/kg), and the impedance measurements were made every 12 s and averaged over 2 min. A 5 min washout period was included between each MCh challenge that included 2 deep inspirations. The cardiovascular function of the mice was inferred from heart rate data acquired using electrocardiogram electrodes. To test the relaxation effect of albuterol or PGE₂, the mice were preconstricted with a dose of MCh that produced an equivalent increase in resistance [wild-type (WT) at 125 μg/kg and βarr-1^−/− at 200 μg/kg i.v.]. Albuterol (15, 60, or 240 µg/kg) or PGE₂ (5, 20, 80 μg/kg) was administered with a 5 min recovery between increasing doses. The response to MCh in the presence of albuterol or PGE₂ was compared with the contractile response to MCh alone, and the percent change was calculated and used to determine the net decrease in the MCh-induced airway response.

Ex vivo analysis of ASM tension development

Murine tracheae excised after euthanasia were cleaned of surrounding connective tissue and mounted into a multiwire myograph chamber in Krebs-Henseleit solution (pH 7.40–7.45) maintained at 37°C with 5% CO₂ and 95% O₂, with frequent changing of the solution. The chambers were mounted onto the Myo-Interface (model 610 M) and connected to the transducer and PowerLab (ADInstruments, Colorado Springs, CO, USA) for data transferring and recording. Chart5 software for Windows (ADInstruments) was used to collect and analyze the data. A basal tension of 0.5 g was set and the tracheal rings were stimulated with increasing concentrations of MCh (10 nM to 100 µM) for 5 min. The rings were washed and preload reset to 0.5 g for 30 min, then stimulated with 60 mM KCl. After stabilization the tracheal rings were again washed, preload reset to 0.5 g, then stimulated with 10 µM MCh and allowed to contract for 10 min. Stimulation of the rings with 10 µM MCh produces ∼ 80% maximum tension in the rings as determined previously (22). Subsequently the precontracted rings were stimulated with increasing concentrations of isoproterenol (ISO; 1 nM to 100 µM) for 5 min, tension recorded, and maximal inhibition calculated as per Tilley et al. (23).

Generation of murine and human ASM cultures

For generation of murine ASM cultures, tracheae harvested from 5 mice were cleaned of surrounding tissues and pooled to isolate ASM cells using the procedure for human ASM cultures with minor modifications, as described previously (24). Human ASM cultures were established as described previously (25) from human airways obtained from lung transplant donors, under procedures approved by the University of Maryland Medical Center and Thomas Jefferson University Institutional Review Board. Characterization of these cells with regard to immunofluorescence of smooth muscle actin and agonist-induced changes in cytosolic calcium has been reported previously (26). Third to sixth passage cells were plated at a density of 10⁴ cells/cm² and maintained in Ham’s F-12 medium supplemented with 10% fetal bovine serum. Cells were growth-arrested 24 or 48 h prior to stimulation by washing once in PBS and refeeding with Ham’s F-12 media supplemented with 5 µg/ml each of insulin and transferrin as per Penn et al. (27). For select experiments, human ASM lines with stable expression of human telomerase reverse transcriptase (hTERT; a generous gift from Bill Gerthoffer, University of South Alabama, Mobile, AL, USA) were employed and maintained as described elsewhere (28).

Small interfering RNA–mediated knockdown of βarr isoforms in ASM

Small interfering RNA (siRNA) ON-TARGETplus SMARTpool oligos (Dharmacon, Lafayette, CO, USA) directed against βarr-1 or βarr-2 were used. Control (scrambled) siRNA oligos were annealed at 37°C for 1 h; 5 μg of the annealed oligo mixture was used to transfect human ASM cells using Dharmafect 1 (Dharmacon, Lafayette, CO, USA) as per manufacturer’s instruction. Twenty-four hours after transfection, the cells were replated for subsequent assessment of βarr protein expression by immunoblot analysis or cAMP accumulation and [Ca²⁺]_i measurement. Additional experiments comparing sham (no siRNA)-transfected cells with βarr-1 or βarr-2 siRNA-transfected cells produced qualitatively similar results to experiments comparing control scrambled siRNA- with βarr-1 or βarr-2 siRNA-transfected cells (data not shown).

cAMP accumulation in cultured ASM cells

Cells transfected with scrambled, βarr-1, or βarr-2 siRNA were plated into 24 well plates and analyzed for agonist-induced cAMP accumulation as described previously (27). Briefly, cells were growth-arrested, washed, and stimulated with PBS containing 300 µM ascorbic acid, 1 mM 3-isobutyl-1-methylxanthine, and either vehicle, ISO (1 nM–10 µM), 0.01 or 1 µM PGE₂, or 100 µM forskolin (FSK), for 10 min at 37°C. Reactions were quenched by buffer aspiration and addition of 400 µl cold 100% ethanol. Isolated cAMP was subsequently quantified by radioimmunoassay (27). FSK-stimulated cAMP was not affected by arrestin subtype KO or knockdown, and agonist-stimulated cAMP production was normalized to FSK-stimulated values as per elsewhere (29) to minimize group variability attributed to modest differences in cell density per well.

Intracellular calcium mobilization

Cells were plated onto collagen-coated 96 well plates, grown to confluence and loaded with 2 µM Fluo-4 AM (BD Biosciences, San Jose, CA, USA) and probenecid (organic anion transporter inhibitor to aid the retention of Fluo-4 in cells) for 1 h. Indicated agonists were added by an automated pipetting system in duplicate, and the 525 nm signals were generated by excitation at 485 nm with a Flex Station II (Molecular Devices, Sunnyvale, CA, USA) as per Deshpande et al. (30). The net peak Ca²⁺ response was calculated as [(agonist-induced fluorescence units) − (vehicle-induced fluorescence units)]. Maximal Ca²⁺ response was determined by stimulating the cells with ionomycin (1 µM).

Radioligand binding

Freshly collected lung and epithelium denuded tracheae and bronchi were homogenized and the membrane pellet prepared (16) and frozen until binding assay. Protein concentration was measured using Bradford reagent. Density of βAR and mAChR expressed on lung and ASM membranes were measured by radioligand binding assays. Pilot study was conducted to select the appropriate concentrations of ligands and assay conditions. Total βAR binding was determined by incubating the membranes with a saturating concentration (500 pM) of [¹²⁵I]-cyanopindolol (a nonselective βAR antagonist; Perkin Elmer, Waltham, MA, USA) for 90 min at room temperature to achieve a steady state. Propranolol (10 µM), a nonselective blocker of βAR was used to determine nonspecific binding. Specific binding was obtained from the difference between total and nonspecific binding. After the incubation, membranes were harvested onto Whatman GF/B glass filters and the radioactivity was measured using a γ counter. Total mAChR binding was determined by incubating the membranes with a saturating concentration (500 pM) of ³H-N-methylscopolamine (a muscarinic receptor antagonist; Perkin Elmer) for 120 min at room temperature to achieve a steady state. Atropine (1 µM), a nonselective blocker of mAChR was used to determine nonspecific binding. After the incubation, membranes were harvested onto Whatman GF/B glass filters, soaked overnight in scintillating solution and the radioactivity was measured using a liquid scintillation analyzer.

Statistical analysis

Data analysis was performed using GraphPad Prism (San Diego, CA, USA) or SPSS (IBM SPSS, Chicago, IL, USA), and data are expressed as means ± se. For in vitro and ex vivo experiments, group comparisons were performed using 1-way ANOVA followed by Newman-Keuls multiple comparison test or a 2-way ANOVA followed by Bonferroni’s multiple comparison test. For in vivo experiments 2-tailed, unpaired Student’s t test was used between different genotypes; repeated-measures ANOVA followed by Bonferroni’s multiple comparison test was used to analyze differences between genotypes with different doses. P < 0.05 was considered sufficient to reject the null hypothesis.

RESULTS

KO or knockdown of βarr-2, but not of βarr-1, augments βAR-stimulated cAMP production

The absence of βarr-1 and βarr-2 protein in the respective KO mice was confirmed by Western blotting in lysates of whole lung homogenates, tracheal homogenates, and isolated tracheal smooth muscle cells (Fig. 1). In addition, siRNA-mediated knockdown of βarr-1 and βarr-2 resulted in a 76.8 ± 2.5 and 74.0 ± 4.2% decrease, respectively, in expression of the βarr isoforms in human ASM cells.

Figure 1.

βarr expression in murine tissues and in cultured murine and human ASM cells. A) Immunoblot analysis of βarr-1 and βarr-2 expression in mouse lung, trachea, and cultured murine ASM cells from WT, βarr-1-KO, and βarr-2-KO mice. Arrestin expression was detected using the A1CT antibody. B) Immunoblot analysis of βarr-1 and βarr-2 expression in human ASM cultures treated with scrambled (Scr), βarr-1 or βarr-2 siRNA. IB, primary antibody used for immunoblotting. Values are means ± sem from 9 experiments.

KO of βarr-2 increased ISO-induced cAMP production in murine ASM cells by 47% whereas βarr-1 KO had no effect (Fig. 2A). In human ASM cells, knockdown of βarr-2 resulted in increased maximal ISO-stimulated cAMP production by 60%, with no effect of βarr-1 knockdown (Fig. 2B). PGE₂-induced cAMP production was not significantly affected by the knockdown of either βarr isoform (Fig. 2C). Neither genotype nor knockdown had a statistically significant effect on calculated −logEC₅₀ values in murine (WT: 7.21 ± 0.25; βarr-1^−/−: 7.36 ± 0.30; βarr-2^−/−: 7.16 ± 0.20 M) or human cells [control (scrambled): 6.88 ± 0.02; βarr-1: 6.87 ± 0.04; βarr-2: 6.74 ± 0.03 M], respectively.

Figure 2.

Effects of βarr knockdown on GPCR signaling in murine and human ASM cultures. ASM cultures derived from WT, βarr-1-KO, and βarr-2-KO mice (n = 4 experiments) (A) or human ASM cells (6–8 experiments) (B, C) were stimulated with ISO (10⁻⁹–10⁻⁵ M) or PGE₂ (10⁻⁸ or 10⁻⁶ M). cAMP accumulation was assessed as described in Materials and Methods. D) Contractile agonist-stimulated [Ca²⁺]_i flux in human ASM cells. n = 6 for THR, HIST, and MCh, n = 4 for Sulp and S1P). E) MCh-stimulated [Ca²⁺]_i in hTERT human ASM cells (n = 6). Veh, vehicle; THR, thrombin 1 U/ml; HIST, histamine 10 µM; MCh, methacholine 10 µM; SCR, scrambled; Sulp, sulprostone 1 µM; S1P, sphingosine-1-phosphate 100 µM. Values represent means ± sem. *P < 0.05; **P < 0.01; ***P < 0.001 vs. WT or SCR control; 2-way ANOVA followed by Bonferroni’s multiple comparison test.

βarr-1 knockdown impairs M3 mAChR signaling

To determine the role of βarr-1 and βarr-2 in G_q protein-coupled receptor signaling, intracellular calcium mobilization was assessed in siRNA-transfected human ASM cells. No effect of arrestin knockdown was found with respect to calcium mobilization mediated by thrombin, histamine, MCh, sulprostone, or sphingosine-1-phosphate (Fig. 2D). However, in both ASM cells cultured from human and murine airways, the increase in intracellular calcium responses to the mAChR receptor agonist MCh was relatively low in our studies, which is consistent with the waning of M3 mAChR signaling in cultured ASM cells (31, 32). To exclude the potential confounding effects of the waning of M3 mAChR expression in ASM cells (the M3 mAChR subtype mediates the calcium response to mAChR agonists in ASM), we used hTERT immortalized human ASM cells, which have previously been shown to retain M3 mAChR receptor signaling (33). The MCh-stimulated [Ca²⁺]_i increase in hTERT ASM cells was ∼10-fold higher than that elicited in non-hTERT cells. In hTERT ASM cells βarr-2 knockdown did not have a significant effect on the maximal increase in [Ca²⁺]_i (Fig. 2E). However, βarr-1 knockdown resulted in a 33% decrease (control: 57 ± 4%; βarr-1: 38 ± 6%; βarr-2: 64 ± 7%) of maximal response to MCh (P < 0.05). The −logEC₅₀ for MCh-induced [Ca²⁺]_i increase was not statistically different among groups (control: 6.56 ± 0.03; βarr-1: 6.36 ± 0.02; βarr-2: 6.80 ± 0.02 M).

Arrestins differentially regulate contraction and relaxation of ASM tissue

To address the regulatory role of βarr isoforms on M3 mAChR and βAR regulation of ASM function, myograph analysis of ASM tension development using excised murine tracheal rings was performed. Consistent with the data from isolated cells, E_max for MCh-stimulated contraction of βarr-1^−/− tracheal rings was 24% lower (P < 0.01) than that observed for WT rings (WT: 2.38 ± 0.15; βarr-1^−/−: 1.75 ± 0.22; βarr-2^−/−: 2.12 ± 0.15) (Fig. 3A). MCh-stimulated contraction of βarr-2^−/− tracheal rings was similar to that of WT rings. By contrast, βarr-1^−/− did not affect ISO-stimulated relaxation, whereas βarr-2^−/− augmented the ISO-stimulated maximal relaxation compared with WT airways (P < 0.01) (Fig. 3B). Calculated −logEC₅₀ values for either MCh-stimulated contraction (WT: 5.95 ± 0.22; βarr-1^−/−: -5.82 ± 0.34; βarr-2^−/−: 5.84 ± 0.14 M) or ISO-stimulated relaxation (WT: 6.79 ± 0.20; βarr-1^−/−: 6.62 ± 0.23; βarr-2^−/−: 6.80 ± 0.11 M) were not statistically different among genotypes. In a separate series of experiments, tracheal preparations of WT mice were relaxed with ISO in the presence of the selective β₂AR antagonist ICI 118,551 (1 µM), which inhibited ISO-induced relaxation by 54% at 1 µM ISO (vehicle: 66.6 ± 2.8%; ICI 118,551: 85.5 ± 3.5%; P < 0.001) (Fig. 3C), suggesting that both β₂AR and β₁AR contribute the ISO-mediated relaxation of contracted murine trachea.

Figure 3.

Effects of βarr KO on tracheal ASM contraction and relaxation ex vivo. Tracheal rings excised from WT, βarr-1^−/−, and βarr-2^−/− mice were mounted in a bath and contractile properties were assessed as described in Materials and Methods. A) Active tension development stimulated by MCh (normalized to 60 mM KCl) was compared between WT and βarr^−/− groups (6–9 experiments). B) The relaxant effect of increasing concentrations of ISO on rings precontracted with 10 μM MCh was compared among groups (3–7 experiments). Values are means ± sem. C) Relaxant effect of increasing concentrations of ISO on murine tracheal rings (WT) precontracted with 10 μM MCh in the presence or absence of the selective β₂AR antagonist, ICI 118,551 (1 µM). Values are means ± sem (4–5 experiments). **P < 0.01; 1-way ANOVA followed by a Newman-Keuls multiple comparison test; ***P < 0.001; 2-way ANOVA followed by Bonferroni’s multiple comparison test.

KO of the βarr-1 gene impairs the MCh-induced airway resistance in vivo

To assess the role of βarr-1 on airway function in vivo, βarr-1^−/− mice were challenged with MCh and airway responsiveness was measured as described in Materials and Methods. The baseline total respiratory system response was not different between WT (93.8 ± 6.9 cmH₂O · s, n = 6) and βarr-1^−/− (88.9 ± 3.5 cmH₂O · s, n = 11) mice. However, the increase in airway responsiveness following MCh challenge was significantly lower in βarr-1^−/− compared with WT (Fig. 4).

Figure 4.

Airway response to methacholine in βarr-1 KO mice. Increasing doses of MCh were administered intravenously and mice were recovered between doses. Total respiratory system resistance was measured using APTI (airway pressure time index). Values are means ± sem from n = 6 WT and n = 11 for βarr-1^−/− mice. *P < 0.05; ***P < 0.001; repeated-measures ANOVA followed by Bonferroni’s multiple comparison test for differences at each dose.

To accurately assess the role of βarr-1 in regulating bronchorelaxation in vivo, mice need to be comparably preconstricted. Because βarr-1^−/− mice display a reduced bronchoconstrictor response to MCh compared with WT mice, we calculated the doses of MCh required to induce equivalent increases in lung resistance (Newtonian resistance). Accordingly, WT and βarr-1^−/− mice were bronchoconstricted with 125 and 200 μg/kg i.v. MCh, respectively, prior to administration of bronchodilators albuterol or PGE₂.

Our data show that for the bronchodilator protocols (albuterol and PGE₂), WT and βarr-1^−/− baseline airway resistance and MCh-induced (125 and 200 µg/kg, respectively) airway resistance were not significantly different among genotypes. As shown in Fig. 5, the WT and βarr-1^−/− percentage change in airway resistance to increasing concentrations of albuterol (15, 60, or 240 µg/kg) or PGE₂ (5, 20, 80 μg/kg) were not significantly different. Our previous studies demonstrate that specific βarr-2 gene KO augments albuterol-induced bronchorelaxation in vivo, whereas PGE₂-induced bronchorelaxation is not affected (16).

Figure 5.

Effects of βarr-1 KO on bronchodilatory response in vivo. To measure the effect of albuterol on airway responsiveness (Rn), WT and βarr-1^−/− mice were preconstricted with equally effective (125 µg/kg and 200 µg/kg, respectively) doses of MCh. The bronchoconstrictor response to MCh was the same for WT and βarr-1^−/− mice and was compared with that of combined MCh and albuterol (15, 60, 240 µg/kg) or PGE₂ (5, 20, 80 µg/kg). Values for the albuterol protocol are means ± sem from n = 7 WT and n = 10 for βarr-1^−/−. Analysis by repeated-measures ANOVA failed to discern an effect of genotype (P = 0.732). Values for the PGE₂ protocol are means ± sem from n = 9 WT and n = 7 for βarr-1^−/−. Analysis by repeated-measures ANOVA failed to discern an effect of genotype (P = 0.594).

Ablation of the βarr-1 gene does not affect lung and ASM muscarinic mAChR expression

Because deletion of the βarr-1 gene may influence the expression of the mAChR and this in turn may affect the interpretation of our physiologic data, radioligand binding assay was performed to determine the lung and ASM maximum concentration of receptor binding sites. mAChR densities were not changed by the deletion of βarr-1 (Table 1). WT and βarr-1^−/− mice expressed similar levels of mAChR in whole lung membranes (190 ± 6.7 and 203 ± 10.7 fmol/mg of protein, respectively) and ASM membranes (466 ± 46 and 468 ± 19 fmol/mg of protein, respectively) (Table 1). We previously determined that the deletion of βarr-2 did not affect the expression of the lung and ASM total βAR (16). More recently, we established that the expression and relative proportions of β₁ARs and β₂ARs in murine lung and tracheobronchial smooth muscle are not significantly affected by deletion of βarr-1 or βarr-2 (34).

TABLE 1.

Effect of βarr-1 deletion on lung and ASM mAChR density (B_max) in mice

	WT lung	βarr-1 KO lung	WT ASM	βarr-1 KO ASM
mAChR	190 ± 7	203 ± 11	466 ± 46	468 ± 19

Membranes from C57BL/6J (WT) and βarr-1-KO mice were prepared as described in Materials and Methods. B_max values were measured by radioligand binding assay (fmol/mg of protein) and expressed as mean ± sem (n = 3–5).

DISCUSSION

In this study, we demonstrate for the first time that βarr isoforms differentially regulate GPCR function in ASM. In accordance with our previous findings (16), specific βarr-2 KO or knockdown augmented ISO-induced cAMP accumulation in vitro and ASM relaxation ex vivo and in vivo. By contrast, in the present study we demonstrate that βarr-1 KO or knockdown does not affect cAMP production or bronchorelaxation. A recent study using radioligand binding demonstrates that murine tracheal smooth muscle expresses both β₁AR and β₂AR (34). We found that the selective competitive antagonist of the β₂AR, ICI 118,551, strongly inhibited ISO-induced relaxation of murine tracheal preparations indicating β₂AR-mediated relaxation. In addition, a previous study has shown that in β₂AR-KO mice albuterol does not induce bronchorelaxation in vivo (35). Thus, although both β₁AR and β₂AR contribute to the relaxant effect on murine tracheae, our results clearly demonstrate that β agonist-induced cAMP accumulation and airway relaxation are regulated by βarr-2.

Interestingly, in hTERT ASM cells that retain M3 mAChR signaling, βarr-1 knockdown impairs MCh-stimulated [Ca²⁺]_i influx whereas βarr-2 knockdown does not. Furthermore, MCh-stimulated ASM contraction ex vivo and bronchoconstriction in vivo are impaired in βarr-1^−/− mice but are not affected in βarr-2^−/−. Signaling by other G_q-coupled GPCRs was not affected by βarr-1 or βarr-2 knockdown. Collectively, our data indicate a differential role for the βarr-1 and βarr-2 arrestin isoforms in regulating GPCR function in ASM: βarr-2 selectively desensitizes the β₂AR, whereas βarr-1 may be involved in maintaining M3 mAChR function.

Members of the GPCR superfamily that activate the heterotrimeric G protein G_q are primarily responsible for the procontractile signaling events in ASM, whereas GPCRs that activate G_s promote relaxation. Acetylcholine, released from parasympathetic nerves, is the dominant physiologic regulator of ASM tone through activation of G_q-coupled M3 mAChRs (36, 37). Signaling events mediated by G_q-coupled receptors have been extensively characterized and include activation of phospholipase C and production of inositol 1,4,5-trisphosphate, which binds to inositol 1,4,5-trisphosphate receptors, resulting in increased [Ca²⁺]_i and, ultimately, contraction, as reviewed elsewhere (9, 38). Under pathophysiologic conditions, such as asthma, ASM tone is increased, in part due to exaggerated presentation of G_q-coupled receptor agonists (including acetylcholine, histamine, leukotrienes, serotonin, and some prostanoids) causing increased procontractile signaling. Most bronchorelaxant therapies are therefore aimed at altering GPCR function. Muscarinic acetylcholine and cysteinyl leukotriene receptor antagonists block specific G_q-coupled receptor signaling.

The other most common therapy for acute asthma attacks, inhaled β-agonist, activates ASM β₂ARs to stimulate cAMP production and PKA activity, which antagonizes G_q-coupled receptor signaling at multiple junctures (1, 39–41). Despite their effectiveness, chronic use of β-agonists is associated with a loss of responsiveness to the drug, as reflected in a loss of bronchoprotective effect (4, 5, 42, 43), deterioration of asthma control, and susceptibility to exacerbations (44). The diminished ability of the β₂AR to effectively promote signaling, or the β₂AR desensitization, is presumed to underlie these physiologic and clinical consequences.

GRK-mediated receptor phosphorylation and the subsequent βarr binding are major determinants of β₂AR desensitization and internalization. βarr-1 and βarr-2 are ubiquitous regulators for the GPCR family (45) but despite their widespread expression, they display nonoverlapping functions and receptor specificity (46). How the βarr isoforms, which have high sequence homology (78%), are able to distinguish among and regulate specific GPCRs is not fully understood. However, numerous studies, when taken together, suggest that the unfolding of βarr initiated by the GRK-phosphorylated GPCR C-terminal tail leads to distinct active conformations for βarr-1 and βarr-2 and thus, receptor-βarr specificity of function (47–49). Although βarr-2 appears to be expressed at relatively low levels in most tissues, a biochemical analysis of the kinetics of βarr-1 and βarr-2 binding properties suggests that βarr-2 may be the predominant physiologic regulator of certain GPCRs despite its low level of expression (50).

Indeed, in murine and human ASM cells and tissue βarr-2 is expressed at much lower levels than is βarr-1. However, our findings that βarr-2 (but not βarr-1) KO or knockdown augments β-agonist-induced cAMP accumulation and ASM relaxation ex vivo and in vivo clearly support the idea that βarr-2 is the key regulator of β₂AR signaling and function under physiologic conditions. The importance of βarr-2 in this process is further emphasized by the fact that alternative desensitization mechanisms known to be active in ASM, such as PKA-mediated receptor phosphorylation (51), are not sufficient to compensate for the loss of βarr-2.

Our data indicate that not only is βarr-1 not involved in desensitization of the GPCR receptors studied here, but it enables MCh-induced ex vivo ASM contraction and in vivo increase of airway resistance. Using hTERT-immortalized human ASM cells that retain M3 mAChR signaling, we were also able to discern this regulatory role of βarr-1 on the M3 mAChR in vitro where it otherwise goes unnoticed in cultured ASM cells, which incur a rapid loss M3 mAChR responsiveness with passage.

The observation that the PGE₂-induced cAMP accumulation and bronchorelaxation are not affected by βarr-1 or βarr-2 KO or selective knockdown of either βarr isoform is consistent with previous findings showing a lack of effect of dual βarr-1/2 knockdown or βarr-2 KO on PGE₂-induced responses (16). Our current data provide further support for a lack of agonist-induced, βarr-mediated desensitization of EP₂ receptor. The PGE₂-activated EP₂ (and possibly EP₄) receptor is expressed in ASM; multiple ASM cell and tissue studies demonstrate that PGE₂ is actually more efficacious than is β-agonist in stimulating intracellular cAMP accumulation and PKA activation and in relaxing ASM tissue ex vivo (16, 22, 41). In vivo, the ability of PGE₂ to prevent/reverse bronchoconstriction appears compromised by the existence of other EP receptor subtypes (some serving bronchoconstriction) in the lung, although a strong bronchorelaxant effect of EP₂ receptor activation can be discerned from studies of EP_1–4 receptor KO mice (23). Recent studies have, however, indicated that PGE₂-induced relaxation of human trachea may be mediated by EP₄ receptors (52), urging caution in extrapolating data from mice to humans. Our findings in both human ASM cells and murine ASM cells and tissue indicate that arrestins do not modulate PGE₂-mediated cAMP accumulation and bronchorelaxation.

Binding studies revealed that the deletion of βarr-1 did not affect the lung and ASM density of βAR and mAChR, which were comparable to that of WT mice. Unfortunately, the lack of useful mAChR subtype-selective ligands does not enable a definitive analysis of M3 mAChR subtype receptor density in the limited amount of ASM tissue available. However, quantitative PCR analysis demonstrated no difference in M3 mAChR mRNA abundance between WT and βarr-1^−/− ASM (not shown). Collectively, these findings fail to implicate reduced M3 mAChR expression as a causal factor mediating decreased MCh-induced contraction observed in βarr-1^−/− mice. One potential explanation relates to a role of βarr-1 in regulating subcellular distribution of the M3 mAChR, perhaps in determining the specific plasma membrane localization or regulating recycling kinetics.

The preference of GPCRs for binding one βarr isoform versus the other suggests that βarr-1 and βarr-2 have distinct functions that are receptor-specific. Our current data demonstrate that βarr-2 is the arrestin isoform that specifically desensitizes β₂AR and constrains β-agonist induced signaling and relaxation in ASM. βarr-1, on the other hand, is required to maintain contractile signaling but is not involved in β₂AR desensitization. The study importantly illustrates that GPCR regulation in more physiologic models with endogenous levels of receptor expression may differ substantially from that observed in overexpression systems. Understanding the role played by each βarr isoform in the signaling and desensitization of endogenous ASM receptors offers unique opportunities for advancing therapeutic strategies aimed at balancing pro- and anticontractile forces in asthma.

Glossary

APTI

airway pressure time index

ASM

airway smooth muscle

β₂AR

β₂-adrenoceptor

βarr

β-arrestin

FSK

forskolin

GPCR

G protein–coupled receptor

GRK

G protein-coupled receptor kinase

hTERT

human telomerase reverse transcriptase

ISO

isoproterenol

KO

knockout

mAChR

muscarinic acetylcholine receptor

MCh

methacholine

PGE₂

prostaglandin E₂

PKA

protein kinase A

siRNA

small interfering RNA

WT

wild-type