Β2-Adrenoceptor Agonist-Induced RGS2 Expression Is a Genomic Mechanism of Bronchoprotection That Is Enhanced by Glucocorticoids

β2-Adrenoceptor agonist-induced RGS2 expression is a genomic mechanism of bronchoprotection that is enhanced by glucocorticoids

Neil S. Holdena, Matthew J. Bellb, Christopher F. Ridera, Elizabeth M. Kinga, David D. Gaunta, Richard Leigha, Malcolm Johnsonc, David P. Siderovskid, Scott P. Heximere, Mark A. Giembycza, and Robert Newtona,1

aAirways Inﬂammation Research Group, Faculty of Medicine, University of Calgary, Calgary, AB, Canada T2N 4N1; bRespiratory Inﬂammation and Autoimmunity Department, MedImmune, Cambridge CB21 6GH, United Kingdom; cGlaxoSmithKline Research and Development, Uxbridge, Middlesex UB6 0HE, United Kingdom; dDepartment of Pharmacology, University of North Carolina School of Medicine, Chapel Hill, NC 27599-7365; and eDepartment of Physiology, University of Toronto, Toronto, ON, Canada M5S 1A8

Edited by Joseph A. Beavo, University of Washington School of Medicine, Seattle, WA, and approved October 17, 2011 (received for review June 23, 2011)

In asthma and chronic obstructive pulmonary disease, activation of signal to phospholipase Cβ via the heterotrimeric G protein, Gq (6, Gq-protein–coupled receptors causes bronchoconstriction. In each 7). This leads to the formation of inositol(1,4,5)trisphosphate 2+ case, the management of moderate-to-severe disease uses inhaled (IP3), which stimulates the release of intracellular free Ca β 2+ corticosteroid (glucocorticoid)/long-acting 2-adrenoceptor agonist ([Ca ]i) from internal stores, and diacylglycerol, an activator of fi (LABA) combination therapies, which are more ef cacious than ei- protein kinase C, which, together with IP3, promotes ASM con- ther monotherapy alone. In primary human airway smooth muscle traction (6, 7). Conversely, β2-adrenoceptor agonists are bron- cells, glucocorticoid/LABA combinations synergistically induce the chodilators that functionally antagonize ASM contraction expression of regulator of G-protein signaling 2 (RGS2), a GTPase- irrespective of the constrictor (8). Current dogma holds that β2- activating protein that attenuates Gq signaling. Functionally, RGS2 adrenoceptor agonism activates the cAMP/cAMP-dependent fl reduced intracellular free calcium ux elicited by histamine, meth- protein kinase A (PKA) pathway to phosphorylate downstream acholine, leukotrienes, and other spasmogens. Furthermore, protec- targets that oppose actin–myosin cross-bridge formation and re- tion against spasmogen-increased intracellular free calcium, follow- duce ASM force generation (8). MEDICAL SCIENCES ing treatment for 6 h with LABA plus corticosteroid, was dependent Signaling from GPCRs is terminated by GTPase-activating on RGS2. Finally, Rgs2-deficient mice revealed enhanced broncho- proteins (GAPs) (9), which enhance the rate at which Gα subunits constriction to spasmogens and an absence of LABA-induced bron- hydrolyze GTP. We show that the expression of the regulator of G- choprotection. These data identify RGS2 gene expression as a protein signaling 2 (RGS2), a G selective GAP (10, 11), is tran- genomic mechanism of bronchoprotection that is induced by glu- q cocorticoids plus LABAs in human airway smooth muscle and pro- scriptionally up-regulated by LABAs and that this is enhanced and vide a rational explanation for the clinical efficacy of inhaled prolonged by glucocorticoids in human ASM. In vitro RGS2 corticosteroid (glucocorticoid)/LABA combinations in obstructive down-regulates signaling from multiple Gq-coupled receptors and airways diseases. in vivo imparts bronchoprotection against spasmogens. Increased RGS2 gene expression therefore represents a genomic mechanism adrenoreceptor | beta-2-adrenergic receptor | protein kinase A | of bronchoprotection, which may contribute to the therapeutic glucocorticoid receptor | NR3C1 activity of ICS/LABA combination therapies. Results sthma affects up to 300 million people globally, and chronic RGS2 Expression Is Induced by LABAs and Enhanced by Glucocorticoids. obstructive pulmonary disease (COPD), which is predomi- A Primary human ASM cells were treated with maximally effective nantly caused by cigarette smoking, is predicted to become the third leading cause of death. Both diseases are characterized by concentrations of dexamethasone, the LABAs, salmeterol xina- airflow limitation due, at least in part, to bronchoconstriction. foate (salmeterol), formoterol fumarate (formoterol), and the Therapeutically, long-acting β -adrenoceptor agonists (LABAs), adenylyl cyclase activator, forskolin, before analysis for RGS2 2 mRNA (Fig. 1A and Fig. S1A) (12). Although RGS2 mRNA was short-acting β2-adrenoceptor agonists (SABAs), and inhaled glucocorticoids, clinically known as inhaled corticosteroids (ICSs), modestly induced by dexamethasone at all times, the LABAs fi represent cornerstone treatments for both asthma and COPD (1, signi cantly induced RGS2 expression at 1 and 2 h, and forskolin 2). Whereas SABAs provide short-term symptomatic relief of induced RGS2 mRNA at all times tested. In the presence of fi bronchospasm, international practice guidelines recommend ICS/ LABAs or forskolin, dexamethasone signi cantly enhanced RGS2 LABA combination inhalers for asthma patients whose symptoms expression at 1 and 2 h. This effect was more than the sum of the are not adequately controlled on ICS monotherapy (1). Moreover, components and is therefore described as synergy. At 6 h, neither fi clinical guidelines recommend long-acting bronchodilators as a LABAs nor dexamethasone signi cantly increased RGS2 mRNA, fi first-line therapy for mild COPD and ICS/LABA combinations for yet the combination signi cantly induced RGS2 mRNA. more severe disease (1, 2). In each case, large, multicenter ran- domized controlled clinical trials convincingly show enhanced clinical benefits to ICS/LABA combinations that are not achieved Author contributions: N.S.H., M.J.B., R.L., M.A.G., and R.N. designed research; N.S.H., M.J.B., – fi C.F.R., E.M.K., D.D.G., M.A.G., and R.N. performed research; D.P.S. and S.P.H. contributed by either component alone (2 4). However, despite the bene ts new reagents/analytic tools; N.S.H., M.J.B., C.F.R., E.M.K., D.D.G., M.A.G., and R.N. ana- achieved by these combination therapies, there is a paucity of lyzed data; and N.S.H., M.J.B., C.F.R., R.L., M.J., D.P.S., S.P.H., M.A.G., and R.N. wrote mechanistic information accounting for these effects. the paper. Several mediators, including acetylcholine, histamine, cysteinyl- The authors declare no conflict of interest. leukotrienes, and certain prostanoids, promote bronchocon- This article is a PNAS Direct Submission. striction, which is a defining characteristic of asthma and COPD (1, 1To whom correspondence should be addressed. E-mail: [email protected]. 2, 5). Such mediators contract airway smooth muscle (ASM) by This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. activating G-protein–coupled receptors (GPCRs) that typically 1073/pnas.1110226108/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1110226108 PNAS Early Edition | 1of6 Downloaded by guest on September 27, 2021 ABthe presence of glucocorticoid are primarily due to the modulation of gene transcription. RGS2 protein expression revealed variable and barely detec- tible increases by the glucocorticoids, fluticasone propionate (FP), budesonide, and dexamethasone (Fig. 1C; Fig. S2 A and B; Fig. S3A). Conversely, salmeterol and formoterol strongly induced RGS2 expression, which peaked at 2 h (Fig. 1C; Figs. S2 and S3). This was mimicked by forskolin, but the duration of RGS2 expression was prolonged. The presence of FP further elevated sal- C meterol- and forskolin-induced RGS2 expression at 1 and 2 h, and, at times thereafter, RGS2 expression was dramatically prolonged (Fig. 1C). Indeed, salmeterol plus FP induced RGS2 protein from 45 min to 18 h posttreatment, with a plateau of near maximal expression between 2 and 8 h (Fig. 1D and Fig. S2). D E Similar data were described for salmeterol plus dexamethasone and formoterol plus budesonide (Figs. S2 and S3). The up-regulation of RGS2 expression by salmeterol or by salmeterol plus FP was attenuated by the β2-adrenoceptor selective antagonist, ICI fi β Fig. 1. Expression of RGS2 in human ASM cells. (A) Primary human ASM 118,551 (Fig. S3C). This con rms the selectivity for the 2 adre- cells were treated with combinations of 1 μM dexamethasone (Dex) or 100 noceptors. Similarly, the ability of salmeterol or salmeterol plus nM salmeterol (Salm). Cells were harvested for real-time PCR analysis of dexamethasone to induce RGS2 expression was totally prevented RGS2 and GAPDH mRNAs (Upper) or unspliced RGS2 nuclear RNA (nRGS2) by the RNA polymerase II inhibitor, actinomycin D, and is con- and U6 RNA (Lower). (Upper) Data (n = 9), expressed as RGS2/GAPDH, are sistent with transcriptional control (Fig. S3D). plotted as fold relative to untreated (at t = 1 h) as means ± SE. (Lower) Data To explore the role of PKA, cells were infected with adenovirus (n = 10) are plotted as nRGS2/U6 as means ± SE. Multiple comparisons at α fi overexpressing PKI , a highly selective and potent inhibitor of each time were made using ANOVA with a Bonferroni post test. Signi cance PKA (16). Using this method, virtually all ASM cells express PKIα, data relative to untreated samples are *P < 0.05, **P < 0.01, and ***P < 0.001. Significance data relative to Dex-treated samples are aP < 0.05, bP < and this inhibits cAMP response element (CRE) binding protein 0.01, and cP < 0.001. Significance data relative to Salm-treated samples are (CREB) phosphorylation and CRE-dependent transcription (17, dP < 0.05, and eP < 0.01. (B) Primary human ASM cells were not stimulated 18). Induction of RGS2 protein by salmeterol was blocked by Ad5- (NS) or treated with Dex (1 μM), Salm (100 nM), or Dex (1 μM) + Salm (100 PKIα, but not by the control Ad5-lacZ (Fig. 1E). In the presence of nM). After 2 h, actinomycin D (10 μg/mL) was added and RNA was harvested salmeterol plus FP, RGS2 protein was strongly induced and was at 15, 30, 60, 120, and 240 min for RT-PCR analysis of RGS2 and GAPDH. Data also inhibited by Ad-PKIα, but not by Ad5-LacZ (P < 0.001 rel- (n = 7 for 15, 30, and 60 min; n = 3 for 120 and 240 min) expressed as RGS2/ ative to either naive or Ad5-LacZ using ANOVA with a Bonfer- GAPDH are plotted as a percentage of their respective treatment at t =0 roni post test). These data support a PKA-dependent mechanism (i.e., when actinomycin D was added) as means ± SE. (C) Cells as in A were treated with 300 nM fluticasone propionate (FP), 100 nM salmeterol (Salm), for the induction of RGS2 by LABAs. μ and 10 M forskolin (Forsk) before Western blot analysis of RGS2 and 2+ Repression of [Ca ]i Flux by LABAs Is Prolonged by Glucocorticoid. In GAPDH. Blots representative of four such experiments are shown. (D) Pri- 2+ mary human ASM cells were treated with 100 nM salmeterol plus 300 nM human ASM, histamine increased [Ca ]i with near maximal fluticasone propionate (Salm + FP) before Western blotting for RGS2 and responses at 10 μM (Fig. 2A). This was blocked by mepyramine, a GAPDH. Blots representative of four experiments are shown. (E) Primary histamine H1 receptor antagonist (19), and 2-aminoethyldiphenyl human ASM cells were left naive or infected with Ad5-PKI or Ad5-LacZ. After borate (2-APB), an IP3R antagonist (20) (Fig. 2B). Because the H1 24 h, cells were treated with 300 nM fluticasone propionate (FP) and/or 100 receptor couples to Gq (19), this confirms a Gq/IP3/IP3R pathway nM salmeterol (Salm) for 6 h before Western blot analysis for RGS2 and 2+ linked to [Ca ]i flux. Similarly, the acetylcholine analog, meth- GAPDH. Blots representative of three such experiments are shown. Densi- acholine, contracts ASM via the Gq-linked muscarinic M3 re- tometric data for C and D are provided in Fig. S2. 2+ ceptor (6, 19). In the current study, methacholine-induced [Ca ]i was maximal at 30 μM and was blocked by 2-APB (Fig. S4 A and To explore the mechanistic basis for this effect, unspliced nu- B). Likewise, the thromboxane analog, U46619, which is selective – for the Gq-linked thromboxane receptor (TP) (19), maximally in- clear RNA (nRNA) corresponding to the exon 1 intron A and 2+ μ exon 4–intron D splice junctions in RGS2 was examined as an creased [Ca ]i at 1 M and this was blocked by 2-APB, again supporting a conventional G -linked pathway leading to elevated index of transcription rate (Fig S1B). Because splicing reactions q [Ca2+] (Fig. S4 C and D). are not limiting and occur very rapidly (13), the transient accu- i Because RGS2 expression was maximal 2 h after LABA addi- mulation of unspliced nRNA is a surrogate of transcription rate – tion and because this was prolonged to at least 6 h by glucocorti- (14, 15). TaqMan PCR for the RGS2 exon 1 intron A slice junc- coid, the protective role of these treatments was tested on tion revealed a very rapid increase induced by salmeterol that was 2+ spasmogen-increased [Ca ]i. Following a 2-h pretreatment, sal- further enhanced by dexamethasone (Fig. 1A). This effect had meterol and salmeterol plus dexamethasone, but not dexametha- declined by 6 h, but in the presence of dexamethasone was still sone alone, significantly attenuated histamine-, methacholine-, or fi signi cantly induced. An identical effect was observed for the U46619-induced [Ca2+] (Fig. 2C). However, after a 6-h pre- – i RGS2 exon 4 intron D junction (Fig. S1C). treatment, only salmeterol plus dexamethasone significantly attenu- To examine a role for mRNA stability, ASM cells were tested in ated histamine-, methacholine-, and U46619-induced increases in standard actinomycin D chase experiments. Following either no 2+ [Ca ]i relative to control, salmeterol, or dexamethasone alone treatment or stimulation with dexamethasone, salmeterol, or sal- (Fig. 2C). To explore this effect, ASM cells were pretreated with meterol plus dexamethasone for 2 h, actinomycin D was added, dexamethasone, salmeterol, or salmeterol plus dexamethasone and decay of RGS2 mRNA was monitored by real-time RT-PCR for various times before stimulation with histamine (Fig. 2D). fi 2+ (Fig. 1B). No signi cant changes in RSG2 mRNA stability were Whereas dexamethasone failed to alter histamine-induced [Ca ]i, 2+ detected between any treatment group. Taken with the nRNA salmeterol significantly repressed [Ca ]i following 1-, 2-, and 4-h data, this clearly shows the increased RGS2 mRNA expression due pretreatments, and salmeterol plus dexamethasone provided sig- to LABAs and that the enhancement and prolonged expression in nificant protection with 1-to 12-h pretreatments (Fig. 2D). This

2of6 | www.pnas.org/cgi/doi/10.1073/pnas.1110226108 Holden et al. Downloaded by guest on September 27, 2021 ABGene silencing by two targeting, but not control, siRNAs ro- bustly reduced salmeterol plus dexamethasone-induced RGS2 expression (Fig. 3D). Following a 6-h incubation with salmeterol 2+ plus dexamethasone, repression of histamine-increased [Ca ]i was evident in naive and control siRNA-treated cells (Fig. 3E and Fig. S4G). However, the two RGS2-targeting siRNAs prevented this, and equivalent effects were shown for methacholine and U46619 (Fig. 3E and Fig. S4G). Conversely, 2 h of salmeterol plus 2+ dexamethasone repressed agonist-increased [Ca ]i, but this was C unaffected by RGS2-targeting siRNAs (Fig. S4G). Because RGS2 is strongly expressed at both times (Fig. 1 and Fig. S3), we propose that RGS2 is an essential and primary effector of repression at 6 h, but that non-RGS2 repressive mechanisms exist with shorter treatments.

RGS2 Is Bronchoprotective. Given a causal relationship between 2+ elevated [Ca ]i and ASM contraction (6, 7), the above data imply that RGS2 prevents ASM contraction. This was tested in − − an experimental model using Rgs2-deﬁcient (Rgs2 / ) C57/Bl6 mice (Fig. S5A) (21). Following PCR genotyping (Fig. S5 A and B), tracheas were removed for analysis of isometric force gen- erated by the trachealis muscle (Fig. 4A). Methacholine con- D centration dependently increased tension in wild-type animals with a peak of 294 ± 22% compared with the effect of 100 mM − − KCl. With trachealis from Rgs2 / littermates, methacholine achieved greater peak tensions (523 ± 104%) (Fig. 4A). Rgs2 − − protein was expressed in wild type, but not in Rgs2 / tracheas (Fig. 4A, Lower).

− − − MEDICAL SCIENCES 2+ +/ / Fig. 2. Effect of dexamethasone and salmeterol on agonist-enhanced [Ca ]i Lung function of wild-type, Rgs2 , and Rgs2 animals was in human ASM cells. (A) Primary ASM cells loaded with Fluo4 were stimu- compared using the SciReq Flexivent system (22). Lung re- lated with various concentrations of histamine. Representative traces sistance and compliance were measured at baseline, following fl fl showing uorescence/ uorescence at time 0 (F/F0) are plotted. Peak F/F0 inhalation of nebulized PBS, methacholine, or 5-hydroxytrypta- from four such experiments are plotted as means ± SE. (B) Cells were treated mine (5-HT) (Fig. 4B). No differences in baseline resistance, or μ as in A except that 50 nM mepyramine (mep) or 10 M 2-APB were added 30 following PBS inhalation, were detected between wild-type, het- min before a 10-μM histamine challenge. Data (mep, n = 6; 2-APB, n =8)are erozygote, or knockout animals. Increasing methacholine doses plotted as means of F/F0 ± SE. Significance was tested using a Student’s t test < < modestly increased resistance in wild-type animals, and this was where *P 0.05 and ***P 0.001. (C) Cells were treated as in A, except that −/− 100 nM salmeterol (Salm) and/or 1 μM dexamethasone (Dex) were added 2 significantly enhanced in Rgs2 animals (Fig. 4B). Hetero- or 6 h before 10 μM histamine, 30 μM methacholine (MCh), or a 1-μM zygotes revealed an intermediate phenotype. Similar data were U46619 challenge. Data (2 h pretreatment, n = 4; 6 h pretreatment, n =8–10) obtained for 5-HT except that knockout animals showed elevated ± expressed as peak F/F0 are plotted as means SE. Multiple comparisons were resistance at all 5-HT doses (Fig. 4B). Lung compliance (inverse made using ANOVA with a Bonferroni post test. Significance data relative to of stiffness) was significantly lower in knockout and heterozygote a b c untreated samples are P < 0.05, P < 0.01, and P < 0.001. Significance data animals compared with wild-type counterparts (Fig. 4B). In re- d < e < f < relative to Dex-treated samples are P 0.05, P 0.01, and P 0.001. sponse to inhaled methacholine or 5-HT, all animals showed re- Significance between other groups is indicated where ***P < 0.001. (D) Cells were pretreated with Dex (1 μM), Salm (100 nM), or Dex (1 μM) + Salm (100 duced compliance. Knockout animals exhibited the lowest nM) for the indicated times before challenge with histamine (10 μM) and compliance, and heterozygotes were an intermediate phenotype. fi measurement of fluorescence. Peak F/Fo is plotted as means ± SE. Signifi- Graded expression of Rgs2 was con rmed (Fig. S5C). Because cance relative to F/Fo induced by histamine was tested by ANOVA with pathology may affect lung function, lungs were removed for his- a Dunnett’s post test. **P < 0.01; ***P < 0.001. tological examination by hematoxylin and eosin staining (for inflammation), periodic acid Schiff staining (for mucus), and dif- ferential cell counting in bronchoalveolar lavage fluid. No dif- − − − illustrates the enhanced protective effect afforded by the LABA/ ferences were detected between wild-type, Rgs2+/ , and Rgs2 / glucocorticoid combination. mice (Fig. S6).

2+ In cultured mouse tracheal cells (MTCs) expressing the smooth RGS2 Attenuates Agonist-Induced Increases in [Ca ]i. Following in- muscle markers α-smooth muscle actin and calponin (Fig. S7A), fection of primary human ASM cells with RGS2-expressing ade- salmeterol plus dexamethasone induced Rgs2 expression and re- novirus, RGS2 expression correlated with multiplicity of infection duced methacholine-, U46619-, and 5-HT–induced increases in (MOI) and reduced [Ca2+] flux elicited by histamine (Fig. 3 A and 2+ −/− i [Ca ]i (Fig. S7 B and C). Conversely, in Rgs2 MTCs, there was B). At all MOIs, repression was significant relative to Ad5-LacZ no Rgs2 and no protection by salmeterol plus dexamethasone 2+ and was reproduced for methacholine and U46619 (Fig. 3B and against agonist-increased [Ca ]i (Fig. S7 B and C). Although this Fig. S4E). Similarly, bradykinin, LTB4,LTD4, and misoprostol, establishes a functional role for Rgs2 in MTCs, regulation of which act on the Gq-linked B1 and B2,BLT1 and BLT2, CysLT1 mouse Rgs2 differed from human ASM cells. Whereas salmeterol and CysLT2,andEP3 receptors, respectively (19), all elicited in- induced Rgs2 protein with similar kinetics to human ASM and via 2+ fi creased [Ca ]i that was signi cantly repressed by Ad5-RGS2, but the β2-adrenoceptor (Fig. S7 D and E), there was no dexametha- not by LacZ-expressing virus (Fig. S4F). Importantly, KCl-induced sone enhancement in MTCs. Therefore, LABA alone was used 2+ [Ca ]i, a receptor-independent stimulus, was unaffected by in vivo to induce Rgs2 expression. Ad5-RGS2 (Fig. 3C). Thus, RGS2 attenuates procontractile Two hours following intratracheal (i.t.) administration of aero- responses produced by Gq-linked GPCR agonists in primary hu- solized formoterol, there was a dose-dependent protection against man ASM cells. methacholine-induced resistance and loss of compliance that

Holden et al. PNAS Early Edition | 3of6 Downloaded by guest on September 27, 2021 AB C

D E

2+ Fig. 3. RGS2 attenuates agonist-induced [Ca ]i primary human ASM cells. Primary human ASM cells were infected with the indicated MOIs of Ad5-RGS2.HA or Ad5-LacZ. (A) After 24 h, cells were harvested for Western blot analysis of HA-tagged RGS2. Blots are representative of four such experiments. (B) Following

overnight incubation in serum-free media, cells were loaded with Fluo4 before stimulation with 10 μM histamine. Representative traces of F/Fo are shown and peak F/Fo data from four experiments are plotted as means ± SE. (C) ASM cells were infected with 100 MOI of Ad5-RGS2.HA or Ad5-LacZ or left naive. As in B, cells were challenged with 60 mM potassium chloride (KCl). Peak F/Fo (n = 4) is plotted as means ± SE. Comparisons relative to control (Ad5-LacZ) were made by Student’s t test. (D) Primary human ASM cells were transfected with the indicated siRNAs before treatment with 100 nM salmeterol plus 1 μM dexamethasone (Salm + Dex) for 6 h. Western blotting for RSG2 and GAPDH was performed. Blots representative of ﬁve experiments are shown. (E) Alternatively, cells were

loaded with Fluo4 before a 10-μM histamine challenge. Representative traces of F/Fo from nine such experiments are shown (Fig. S4G).

− − correlated with Rgs2 expression (Fig. S8). Wild-type and Rgs2 / against methacholine-induced increases in lung resistance and animals were therefore treated with i.t. formoterol for 2 h before loss of compliance. lung function analysis. Methacholine increased lung resistance and reduced compliance in wild-type animals, and formoterol sig- Discussion − − niﬁcantly protected against these effects (Fig. 4C). In Rgs2 / We describe the induction of RGS2 expression as a previously animals, the changes in lung function induced by methacholine unrecognized genomic mechanism of bronchoprotection that is were unaffected by formoterol. Basal and formoterol-induced induced by LABAs and that is synergistically enhanced by gluco- Rgs2 expression was conﬁrmed in the lungs of wild-type, but not corticoids in human ASM cells. We propose that this represents an − − of Rgs2 / , animals (Fig. 4C). Thus, elevated RGS2 protects important means by which ICS/LABA therapies effect superior

A B

Fig. 4. Rgs2 is bronchoprotective. (A) Equilibrated strips of − − mouse trachealis smooth muscle from wild type (WT) or Rgs2 / knockout (KO) animals were challenged with KCl (100 mM). At peak contraction, muscle strips were washed free of KCl and re- equilibrated. Cumulative concentration-response curves (n = 11–15) were constructed to methacholine (MCh) and plotted as a percentage of the KCl-induced response as means ± SE. Trachealis muscles were subsequently Western blotted for Rgs2 and Gapdh. Data from three animal pairings are shown. Comparisons were made using a Mann–Whitney U test. Sig- niﬁcance relative to the relevant wild-type control sample is − *P < 0.05. (B) Wild type (WT), Rgs2 /+ heterozygous (Het), and C Rgs2−/− knockout (KO) mice were subjected to lung function analysis. Total lung resistance (Upper panels) and compliance (Lower panels) was measured at baseline (Left panels), following inhalation of nebulized PBS and PBS containing increasing concentrations of methacholine (Center panels) or

5-HT (Right panels). Data (n =6–16) are plotted as cm H2O/s/mL (resistance) or ml/cm H2O (compliance) as means ± SE. (C) Wild type (WT) (Left panels) and Rgs2 knockout (KO) (Right panels) animals were treated by i.t. instillation with 50 μLof0.9%sa- line or 0.5 mg/kg formoterol (Form). After 2 h, mice were anesthetized before lung function analysis. Total lung resistance (Upper panels) and compliance (Lower panels) were measured in response to increasing concentrations of nebu-

lized methacholine. Data (n =4–6) are plotted as cm H2O/s/mL or ml/cm H2O as means ± SE. Lungs (n = 4) were subsequently Western blotted for Rgs2 and Gapdh.

4of6 | www.pnas.org/cgi/doi/10.1073/pnas.1110226108 Holden et al. Downloaded by guest on September 27, 2021 control of symptoms in diseases such as asthma or COPD. Thus, structurally distinct GR ligands with differing selectivities for international guidelines for the management of asthma and COPD other nuclear hormone receptors, including the progesterone, indicate stepwise increases in the intensity of pharmacotherapy mineralocorticoid, androgen, and estrogen receptors (4), moving from inhaled monotherapies to inhaled ICS/LABA com- excludes their involvement and supports a role for GR. This binations (1, 2). For example, mild-to-moderate asthmatics are interaction between the glucocorticoid and cAMP pathways generally prescribed regular ICS whereas more severe disease requires formal investigation and should be tested alongside requires ICS/LABA combinations, which enhance the therapeutic other physiological agents, such as parathyroid hormone or effectiveness of the ICS to a level that cannot be achieved by ICS angiotensin II, which also induce RSG2 expression (36, 37). alone (23). Conversely, in COPD, LABAs represent an initial Importantly, knockdown of RGS2 expression in human ASM 2+ maintenance medication, which, in more advanced disease, is in- cells did not prevent attenuation of spasmogen-enhanced [Ca ]i creased to ICS/LABA combinations. Equally, bronchiolitis is a following 2 h of salmeterol plus dexamethasone. Given the pres- major cause of hospitalization in infants of less than 1 y and is poorly ence of RGS2 protein, classical mechanisms of bronchoprotection controlled by conventional monotherapies (24). However, inhaled must operate to create functional redundancy. However, the epinephrine, which will activate β2-adrenoceptors on ASM, com- protective effect of 6 h of salmeterol plus dexamethasone on hu- bined with dexamethasone (epinephrine and dexamethasone were man ASM was totally prevented by RGS2 knockdown. This indi- ineffective alone), significantly reduced hospitalization rates in cates that elevated RGS2 gene expression is the dominant these infants (25). Thus, multiple strands of clinical evidence sup- protective mechanism and clearly illustrates how ICS/LABA port a positive, synergistic interaction between the agonist of the β2- combinations may confer enhanced therapeutic benefit. Indeed, 2+ adrenoceptor and the glucocorticoid receptor (GR). Our current the 12-h protection from spasmogen-induced [Ca ]i afforded by study offers some explanation for this effect (23). salmeterol plus dexamethasone, compared with just 4 h for sal- Mechanistically, β2-adrenoceptor agonists can synergistically meterol alone, further demonstrates the potential advantage due enhance glucocorticoid actions, and this may contribute to the to RGS2. In this context, it is salient to note that the effects due to enhanced efficacy of ICS/LABA combinations (12, 23). However, RGS2 occur on a background of other RGS proteins, including we show RGS2 to be a PKA-dependent effector gene, whose ex- RGS3, -4, and -5, which are all relatively highly expressed in hu- pression in human ASM promotes bronchoprotection and is syn- man ASM cells (38). Indeed, somewhat paradoxically, β-agonists ergistically enhanced by glucocorticoids. Functionally, RGS2 is a reduce RGS5 expression, and this may be expected to reduce some GAP that switches off signaling from Gq-coupled GPCRs (10, 26), of the benefit conferred by elevation of RGS2 (38). including muscarinic M1 and M3 receptors (27, 28), and we con- These data also highlight important differences between mice firm this for multiple other receptors that are relevant to lung and humans. First, human RSG2 is synergistically induced by MEDICAL SCIENCES disease. Importantly, combinations of LABA plus glucocorticoid LABAs and glucocorticoids, whereas mouse Rgs2 is induced only enhance and extend RGS2 expression in human ASM, and loss of by LABAs. Thus, the mouse models Rgs2 function, but not gene RGS2 expression prevents the prolonged protective effect of the regulation. Second, mouse Rgs2 function is nonredundant when 2+ combination against agonist-enhanced [Ca ]i. In vivo, loss of induced by LABA alone. In this regard, there is persuasive evi- RGS2 produces a hyperactive phenotype with increased lung re- dence suggesting that the classical bronchodilatory effects of sistance and reduced compliance. This is consistent with the hy- β-adrenoceptor agonists are mediated via the β1, rather than the pertensive effect of RGS2 loss in the vasculature and confirms β2, subtype in the mouse (39). Thus, β2-adrenoceptor agonists are a bronchoprotective role in the lung (29, 30). Furthermore, the poor bronchodilators in the mouse, and this explains why non- ability to induce RGS2 expression in vivo further protects against selective agonists are more effective and potent bronchodilators spasmogen-increased lung resistance and reductions in compli- than β2-selective agonists (40). Nevertheless, our data clearly show ance. Thus, the extended induction of RSG2 in ASM by ICS/ that β2-adrenoceptor agonists induce the expression of Rgs2 and LABA combination therapies should provide superior protection that this leads to a significant nonredundant bronchoprotective −/− from multiple spasmogens that act on Gq-coupled GPCRs in effect that is absent in Rgs2 animals. asthma, COPD, or bronchiolitis and other diseases. In summary, the induction of RGS2 gene expression provides Induction of RGS2 expression by β2-adrenoceptor agonists and bronchoprotection, which, in the context of β2-adrenoceptor forskolin implies a classical cAMP pathway leading rapidly to monotherapy, is likely to be largely redundant because of acute, RGS2 gene expression and is consistent with isoproterenol in- nongenomic mechanisms of relaxation/bronchoprotection. How- ducing RGS2 mRNA in rat astrocytoma cells (31). Roles for the β2- ever, in the clinical context of ICS/LABA combination therapies, adrenoceptor and PKA were confirmed by preventing RGS2 ex- RGS2 expression may be enhanced and prolonged. Under this pression in the presence of ICI 118,551 or PKIα and is supported condition, the genomic induction of RGS2 becomes the pre- for forskolin, which induces CREB binding to CRE sites in the dominant effector for protecting against bronchoconstrictor medi- proximal RGS2 promoter (32). Likewise, the ability of gluco- ators. This represents a fundamental shift in our understanding of corticoids to synergistically enhance RGS2 expression could sim- how sympathomimetic bronchodilators improve symptom control plistically involve enhanced β2-adrenoceptor expression, coupling, in diseases such as asthma, COPD, or acute bronchiolitis. In terms or reduced receptor desensitization (see ref. 8 and references of drug discovery, drugs, or drug combinations, that enhance and/or therein). However, these effects are not likely to occur acutely prolong RGS2 expression are, like ICS/LABA combinations, pre- and in any event may not be important in ASM due to the large β2- dicted to result in superior bronchoprotection from clinically adrenoceptor reserve (33). More likely, as RGS2 is a glucocorti- relevant spasmogens. coid-responsive gene (15, 34) with an established GR-binding region (35), there is a direct interaction of the cAMP and glu- Materials and Methods cocorticoid pathways at the level of promoter activation. Cer- Materials. Reagentswere fromSigma-Aldrichunlessotherwisespecified.C57/Bl6 − − − tainly, our data support a mechanism that involves transcriptional wild-type, Rgs2 heterozygous (Rgs2+/ ), and knockout (Rgs2 / ) mice were bred +/− +/− induction of RGS2 because the level of unspliced nuclear RGS2 in-house through Rgs2 × Rgs2 crosses (21, 30). Protocols were approved by RNA was both enhanced and prolonged in the presence of LABA the University of Calgary Animal Care Committee, according to the Canadian Council for Animal Care guidelines. Mice were age and sex matched between plus glucocorticoid. Because there was no obvious effect on different genotype groups. Where appropriate, mice were anesthetized with RGS2 mRNA stability, and because protein expression correlates isofluorane before i.t administration of aerosolized 0.9% NaCl or formoterol well with mRNA expression, these data support transcriptional using a MicroSprayer aerosolizer (Penn-Century) (41). Isolated ASM cells from control as the major effector mechanism governing RGS2 ex- normal human lungs were grown in Dulbecco’sModified Eagle Medium (Invi- pression in human ASM cells. Furthermore, the use of three trogen) and incubated in serum-free medium before experiments (12).

Holden et al. PNAS Early Edition | 5of6 Downloaded by guest on September 27, 2021 RNA Isolation and Real-Time PCR. RNA extractions and real-time PCR (SYBR of ketamine hydrochloride (90 mg/kg) (Wyeth). Tracheostomized mice were GreenER chemistry) for GAPDH was as described (15). Ampliﬁcation primers connected to a ﬂexivent small animal ventilator (SciReq) before challenge were RGS2 forward (5′-CCT CAA AAG CAA GGA AAA TAT ATA CTG A-3′)and with increasing concentrations of aerosolized drugs administered via an in- reverse (5′-AGT TGT AAA GCA GCC ACT TGT AGC T-3′). TaqMan analysis of line nebulizer. Resistance and compliance were measured using a snapshot unspliced nuclear RGS2 is described in SI Materials and Methods. 150 perturbation (22).

Western Blotting, siRNA Knockdown, and Adenoviral Infections. Western Isometric Force Measurement. Mice were euthanized with 200 mg/kg pen- blotting, siRNA transfections, and adenoviral infections were as described tobarbital and exsanguination. Excised tracheas were divided into four equal ′ (42). siRNA-targeted sequences were RGS2 siRNA 2 (SI00045773; 5 -AAC GTG pieces before placing in oxygenated Krebs–Henseleit solution (KHS) (33). ′ ′ GTG TCT CAC TCT GAA-3 ); RGS2 siRNA 6 (SI03036082; 5 -AAG GGT ATA CAG Strips were mounted vertically under a tension of 6 mN in oxygenated KHS ′ fl CTT GAT GGA-3 ) (Qiagen); and green uorescent protein siRNA (control at 37 °C. After equilibration with frequent washing for at least 120 min – ′ ′ siRNA) (P-002048 03-20; 5 -GGC AAG CTG ACC CTG AAG TTC-3 ) (Dharma- before experiments, changes in tension were measured isometrically (33). con). Cells were infected with 300 MOI of Ad5-RGS2.HA (43) or Ad5-LacZ for 24 h before experiments. ACKNOWLEDGMENTS. This research was supported by grants from the GlaxoSmithKline Collaborative Innovative Research Fund, GlaxoSmithKline 2+ Calcium Assay. [Ca ]i release was measured using Fluo4 NW according to the Canada (to R.N., M.A.G., and R.L.); by operating grants from the Canadian manufacturer’s instructions (Invitrogen). Confluent cells were loaded for 30 Institutes of Health Research (to R.N.); by studentship support from min at room temperature with Fluo-4. Using 494 nm excitation, basal fluo- GlaxoSmithKline, United Kingdom (to R.N. and M.A.G.), studentship awards from the Lung Association of Alberta and North West Territories (to C.F.R.) rescence (F0) was measured at 516 nm for 1.6 s (FLUOstar OPTIMA plate reader, BMG Labtech). Agonists were injected and fluorescence (F) was and from Alberta Innovates-Health Solutions (to C.F.R.); and an Izaak Walton Killam Post-Doctoral Fellowship (to N.S.H.). R.N. and R.L. are Alberta measured. Values expressed as F/F are a proxy for [Ca2+] . 0 i Innovates-Health Solutions Senior Scholars and Clinical Investigators, respectively. A grant from the Canadian Fund for Innovation and the Alberta Lung Function. Mice were anesthetized by intraperitoneal injection of sodium Science and Research Authority provided equipment and infrastructure for pentobarbital (50 mg/kg) (CEVA Santé Animale) and intramuscular injection conducting real-time PCR.

1. Bateman ED, et al. (2008) Global strategy for asthma management and prevention: 22. Shalaby KH, Gold LG, Schuessler TF, Martin JG, Robichaud A (2010) Combined forced GINA executive summary. Eur Respir J 31:143–178. oscillation and forced expiration measurements in mice for the assessment of airway 2. Rabe KF, et al.; Global Initiative for Chronic Obstructive Lung Disease (2007) Global hyperresponsiveness. Respir Res 11:82. strategy for the diagnosis, management, and prevention of chronic obstructive pul- 23. Giembycz MA, Kaur M, Leigh R, Newton R (2008) A Holy Grail of asthma manage- monary disease: GOLD executive summary. Am J Respir Crit Care Med 176:532–555. ment: Toward understanding how long-acting beta(2)-adrenoceptor agonists en- 3. Shrewsbury S, Pyke S, Britton M (2000) Meta-analysis of increased dose of inhaled hance the clinical efficacy of inhaled corticosteroids. Br J Pharmacol 153:1090–1104. steroid or addition of salmeterol in symptomatic asthma (MIASMA). BMJ 320: 24. Wainwright C (2010) Acute viral bronchiolitis in children: A very common condition 1368–1373. with few therapeutic options. Paediatr Respir Rev 11:39–45, quiz 45. 4. Newton R, Leigh R, Giembycz MA (2010) Pharmacological strategies for improving the 25. Plint AC, et al.; Pediatric Emergency Research Canada (PERC) (2009) Epinephrine and efficacy and therapeutic ratio of glucocorticoids in inflammatory lung diseases. dexamethasone in children with bronchiolitis. N Engl J Med 360:2079–2089. Pharmacol Ther 125:286–327. 26. Hains MD, Siderovski DP, Harden TK (2004) Application of RGS box proteins to evaluate – 5. Barnes PJ (2008) Immunology of asthma and chronic obstructive pulmonary disease. G-protein selectivity in receptor-promoted signaling. Methods Enzymol 389:71 88. Nat Rev Immunol 8:183–192. 27. Bernstein LS, et al. (2004) RGS2 binds directly and selectively to the M1 muscarinic 6. Gosens R, Zaagsma J, Meurs H, Halayko AJ (2006) Muscarinic receptor signaling in the acetylcholine receptor third intracellular loop to modulate Gq/11alpha signaling. – pathophysiology of asthma and COPD. Respir Res 7:73. J Biol Chem 279:21248 21256. 7. Mahn K, et al. (2010) Ca(2+) homeostasis and structural and functional remodelling of 28. Tovey SC, Willars GB (2004) Single-cell imaging of intracellular Ca2+ and phospholi- airway smooth muscle in asthma. Thorax 65:547–552. pase C activity reveals that RGS 2, 3, and 4 differentially regulate signaling via the – 8. Giembycz MA, Newton R (2006) Beyond the dogma: Novel beta2-adrenoceptor sig- Galphaq/11-linked muscarinic M3 receptor. Mol Pharmacol 66:1453 1464. nalling in the airways. Eur Respir J 27:1286–1306. 29. Tang KM, et al. (2003) Regulator of G-protein signaling-2 mediates vascular smooth – 9. Siderovski DP, Willard FS (2005) The GAPs, GEFs, and GDIs of heterotrimeric G-protein muscle relaxation and blood pressure. Nat Med 9:1506 1512. alpha subunits. Int J Biol Sci 1:51–66. 30. Heximer SP, et al. (2003) Hypertension and prolonged vasoconstrictor signaling in fi – 10. Heximer SP (2004) RGS2-mediated regulation of Gqalpha. Methods Enzymol 390: RGS2-de cient mice. J Clin Invest 111:445 452. 31. Kim SD, et al. (2006) Mechanism of isoproterenol-induced RGS2 up-regulation in as- 65–82. trocytes. Biochem Biophys Res Commun 349:408–415. 11. Kimple AJ, et al. (2009) Structural determinants of G-protein alpha subunit selectivity 32. Song Y, et al.; CHARGE Consortium; GIANT Consortium (2010) CRTC3 links catechol- by regulator of G-protein signaling 2 (RGS2). J Biol Chem 284:19402–19411. amine signalling to energy balance. Nature 468:933–939. 12. Kaur M, Chivers JE, Giembycz MA, Newton R (2008) Long-acting beta2-adrenoceptor 33. Giembycz MA (2009) An estimation of beta 2-adrenoceptor reserve on human agonists synergistically enhance glucocorticoid-dependent transcription in human bronchial smooth muscle for some sympathomimetic bronchodilators. Br J Pharmacol airway epithelial and smooth muscle cells. Mol Pharmacol 73:203–214. 158:287–299. 13. Audibert A, Weil D, Dautry F (2002) In vivo kinetics of mRNA splicing and transport in 34. Wang JC, et al. (2004) Chromatin immunoprecipitation (ChIP) scanning identifies mammalian cells. Mol Cell Biol 22:6706–6718. primary glucocorticoid receptor target genes. Proc Natl Acad Sci USA 101: 14. Lipson KE, Baserga R (1989) Transcriptional activity of the human thymidine kinase 15603–15608. gene determined by a method using the polymerase chain reaction and an intron- 35. So AY, Cooper SB, Feldman BJ, Manuchehri M, Yamamoto KR (2008) Conservation specific probe. Proc Natl Acad Sci USA 86:9774–9777. analysis predicts in vivo occupancy of glucocorticoid receptor-binding sequences at 15. Chivers JE, et al. (2006) Analysis of the dissociated steroid RU24858 does not exclude glucocorticoid-induced genes. Proc Natl Acad Sci USA 105:5745–5749. a role for inducible genes in the anti-inflammatory actions of glucocorticoids. Mol 36. Grant SL, et al. (2000) Specific regulation of RGS2 messenger RNA by angiotensin II in – Pharmacol 70:2084 2095. cultured vascular smooth muscle cells. Mol Pharmacol 57:460–467. 16. Meja KK, et al. (2004) Adenovirus-mediated delivery and expression of a cAMP-de- 37. Tsingotjidou A, Nervina JM, Pham L, Bezouglaia O, Tetradis S (2002) Parathyroid pendent protein kinase inhibitor gene to BEAS-2B epithelial cells abolishes the anti- hormone induces RGS-2 expression by a cyclic adenosine 3′,5′-monophosphate-me- fl in ammatory effects of rolipram, salbutamol, and prostaglandin E2: A comparison diated pathway in primary neonatal murine osteoblasts. Bone 30:677–684. – with H-89. J Pharmacol Exp Ther 309:833 844. 38. Yang Z, et al. (2011) Beta-agonist-associated reduction in RGS5 expression promotes 17. Clarke DL, Belvisi MG, Hardaker E, Newton R, Giembycz MA (2005) E-ring 8-iso- airway smooth muscle hyper-responsiveness. J Biol Chem 286:11444–11455. prostanes are agonists at EP2- and EP4-prostanoid receptors on human airway 39. Henry PJ, Goldie RG (1990) Beta 1-adrenoceptors mediate smooth muscle relaxation smooth muscle cells and regulate the release of colony-stimulating factors by acti- in mouse isolated trachea. Br J Pharmacol 99:131–135. vating cAMP-dependent protein kinase. Mol Pharmacol 67:383–393. 40. Tomkinson A, Karlsson JA, Raeburn D (1993) Comparison of the effects of selective 18. Kaur M, et al. (2008) Effect of beta2-adrenoceptor agonists and other cAMP-elevating inhibitors of phosphodiesterase types III and IV in airway smooth muscle with dif- agents on inflammatory gene expression in human ASM cells: A role for protein ki- fering beta-adrenoceptor subtypes. Br J Pharmacol 108:57–61. nase A. Am J Physiol Lung Cell Mol Physiol 295:L505–L514. 41. Bivas-Benita M, Zwier R, Junginger HE, Borchard G (2005) Non-invasive pulmonary 19. Alexander APH, Mathie A, Peters JA (2009) Guide to Receptors and Channels (GRAC), aerosol delivery in mice by the endotracheal route. Eur J Pharm Biopharm 61: 4th Ed. Br J Pharmacol 158(Suppl 1):S1–S254. 214–218. 20. Maruyama T, Kanaji T, Nakade S, Kanno T, Mikoshiba K (1997) 2APB, 2-amino- 42. King EM, Holden NS, Gong W, Rider CF, Newton R (2009) Inhibition of NF-kappaB- ethoxydiphenyl borate, a membrane-penetrable modulator of Ins(1,4,5)P3-induced dependent transcription by MKP-1: Transcriptional repression by glucocorticoids oc- Ca2+ release. J Biochem 122:498–505. curring via p38 MAPK. J Biol Chem 284:26803–26815. 21. Oliveira-Dos-Santos AJ, et al. (2000) Regulation of T cell activation, anxiety, and male 43. Salim S, Sinnarajah S, Kehrl JH, Dessauer CW (2003) Identification of RGS2 and type V aggression by RGS2. Proc Natl Acad Sci USA 97:12272–12277. adenylyl cyclase interaction sites. J Biol Chem 278:15842–15849.

6of6 | www.pnas.org/cgi/doi/10.1073/pnas.1110226108 Holden et al. Downloaded by guest on September 27, 2021