Dexamethasone rescues TGF-β1-mediated β2-adrenergic receptor dysfunction and attenuates phosphodiesterase 4D expression in human airway smooth muscle cells

Glucocorticoids (GCs) and β2-adrenergic receptor (β2AR) agonists improve asthma outcomes in most patients. GCs also modulate gene expression in human airway smooth muscle (HASM), thereby attenuating airway inflammation and airway hyperresponsiveness that define asthma. Our previous studies showed that the pro-fibrotic cytokine, transforming growth factor- β1 (TGF-β1) increases phosphodiesterase 4D (PDE4D) expression that attenuates agonist-induced levels of intracellular cAMP. Decreased cAMP levels then diminishes β2 agonist-induced airway relaxation. In the current study, we investigated whether glucocorticoids reverse TGF-β1-effects on β2-agonist-induced bronchodilation and modulate pde4d gene expression in HASM. Dexamethasone (DEX) reversed TGF-β1 effects on cAMP levels induced by isoproterenol (ISO). TGF-β1 also attenuated G protein-dependent responses to cholera toxin (CTX), a Gαs stimulator downstream from the β2AR receptor. Previously, we demonstrated that TGF-β1 treatment increased β2AR phosphorylation to induce hyporesponsiveness to a β2 agonist. Our current data shows that expression of grk2/3, kinases associated with attenuation of β2AR function, are not altered with TGF-β1 stimulation. Interestingly, DEX also attenuated TGF-β1-induced pde4d gene expression. These data suggest that steroids may be an effective therapy for treatment of asthma patients whose disease is primarily driven by elevated TGF-β1 levels.


Background
Asthma, a chronic inflammatory disease of the lungs manifests by several hallmarks: airway hyperresponsiveness, remodeling, and inflammation [1]. Airway smooth muscle (ASM) cells play an integral role in regulating bronchomotor tone in the asthma diatheses and are a direct target of β 2 -agonists, a common therapy that promotes bronchodilation [2]. While β 2 -agonists evoke ASM relaxation, β2-agonists are not effective in all patients [3]. Patients who fall into the "severe" category of asthma are frequently hyporesponsive to bronchodilators [4]. Studies show that β 2 AR tolerance or desensitization occurs after repeated bronchodilator use that diminishes drug efficacy [3,[5][6][7][8][9].
TGF-β1, a profibrotic cytokine whose levels are elevated in patients with asthma, augments human airway smooth muscle (HASM) cell stiffness and significantly increases myosin light chain (MLC) phosphorylation via Smad3 [10] that enhance contractile agonist-induced cell shortening and hyperresponsiveness. In addition to amplifying bronchoconstriction, we also demonstrated that TGF-β1 blunts intracellular cAMP by upregulating pde4d expression that decreases β 2 -agonist-induced cAMP levels [11].
Signaling downstream of seven transmembrane G protein-coupled receptors (GPCRs) involves Gαβγ trimer dissociation following receptor activation [12]. The G α subunit family is comprised of G αi , G αq , and G αs , playing fundamental roles in regulating HASM relaxation and contraction [13]. Cholera toxin (CTX) catalyzes the ADP-ribosylation of G αs that elicits adenylyl cyclase (AC) activation, causing the accumulation of intracellular cAMP and further activation of PKA to induce HASM relaxation [14]. CTX, and the β 2− agonist isoproterenol (ISO), induce actin depolymerization in HASM in PKA-independent and -dependent pathways integrating activation of Src protein tyrosine kinases and G αs protein [15], to promote smooth muscle relaxation. Whether TGF-β1 directly modulates G αs protein activation remains unknown.
Glucocorticoids (GCs) remain a cornerstone in the management of asthma. GC treatment alters gene expression in HASM, thereby modulating inflammation and airway reactivity. Others have demonstrated that dexamethasone, a glucocorticoid, can directly inhibit Smad3 activity [16,17]. Our previous study showed that TGF-β1 blunted the effects of β 2 agonist-induced reversal of carbachol-mediated phosphorylation of myosin light chain, a process which was Smad3-dependent [11]. Given this information, we posited that GC treatment would reverse TGF-β1-induced hyporesponsiveness to β 2 -agonist. Our data demonstrate that dexamethasone (DEX) reverses TGF-β1-induced attenuation of β 2 AR-induced signaling, rescuing β 2 -agonist-and G αs -activatormediated cAMP production by attenuating pde4d expression.

HASM cell culture
HASM cells from the National Disease Research Interchange (Philadelphia, PA, USA) and the International Institute for the Advancement of Medicine (Edison, NJ, USA) were derived from trachea obtained from donors without chronic illness. All tissue was obtained from deidentified donors and was deemed non-human subjects research by the Rutgers University Institutional Review Board. Cells were cultured in Ham's F12 medium with 10% fetal bovine serum. The cells were incubated and grown at 37°C in 5% CO 2 . We have shown that isolated airway smooth muscle cells retain their phenotypic properties [18]. Primary HASM cells between passages 3-4 were used in all experiments. Donor demographics for the cell lines utilized in these studies are detailed in Table S1.

Western blot analysis
Primary HASM cells were serum deprived for 24 h prior to treatment. HASM cells were lysed with 0.6 N HClO 4 , scraped, collected, and pelleted as previously described [19]. The membrane was blocked with ready-to-use Odyssey Blocking Buffer (LI-COR BioSciences) containing 0.1% sodium azide and probed for phospho-Smad3, pMLC, total MLC, and GAPDH.

Measurement of intracellular cAMP levels
Grown to 90% confluency on 24-well plates, HASM cells were stimulated and lysed using the Applied Biosystems cAMP-Screen Immunoassay System following the manufacturer instructions as previously described [11]. The cells were lysed and incubated for 30 min in 5% CO 2 and 37°C. Conjugate Dilution Buffer, cAMP-AP Conjugate, anti-cAMP antibody, and the samples were added to pre-coated assay plate to incubate for 1 h on plate shaker. Plate was measured on luminometer after a 30min incubation period with CSPD®/Sapphire-II RTU Substrate. Data was derived from standard curves and cAMP levels reported after using standard dilutions.
Quantification of pde4d, grk2, and grk3 expression (RNA isolation and qPCR) Following treatment with TGF-β1 ± dexamethasone, cells were suspended in TRIzol reagent, and total RNA were isolated following the manufacturer protocol. RNA was isolated and purified from HASM cells using the RNeasy Mini Kit and cDNA was created using SuperScript IV First-Strand Synthesis System. All reactions were performed in 20 μL reaction volume in triplicate. For mRNA cDNA, PCR amplification consisted of 10 min of an initial denaturation step at 95°C, followed by 40 cycles at 95°C for 15 s, 60°C for 60 s. Relative cDNA quantification was performed using TaqMan quantitative RT-PCR (Thermo Fisher Scientific) and the ΔΔC t method, and pde4d, grk2, and grk3 expression were normalized to expression of endogenous β-actin.

Statistical analysis
Graphs were created and statistical analyses were conducted using GraphPad Prism 5.01 h software (La Jolla, Ca, USA) to determine statistical significance evaluated using two-tailed Student's paired t-test for two groups. P values of < 0.05 were considered significant. All results were confirmed by experiments in at least three unique cell lines. Data was fit to a normal distribution, and appropriate tests run to determine significance. For comparison of multiple conditions, one-way ANOVA was used with Bonferroni's post-test. For the pde4d expression results, the differential expression analysis was performed under a negative binomial distribution model with DESeq2

TGF-β1 attenuates G αs -mediated cAMP production in HASM
Upon activation of β 2 AR, ADP-ribosylation of the α subunit of stimulatory G protein (G αs ), stimulates adenylyl cyclase to increase intracellular cAMP [15,19]. To further understand mechanisms underlying TGF-β1mediated hyporesponsiveness to a β 2 -agonist, intracellular cAMP levels were measured in TGF-β1-treated HASM cells following stimulation with cholera toxin (CTX), a G αs activator. Intracellular cAMP activity increased in a time-dependent manner following exposure to CTX, with the maximum level elicited at 45 min (Fig. 1a). Interestingly at 60 and 75 min, the CTXinduced cAMP levels extinguished. In TGF-β1-treated cells, CTX-induced cAMP levels were completely abrogated compared to that of the diluent control (Fig. 1b).
We previously demonstrated that isoproterenol (ISO) decreased carbachol (Cch)-and TGF-β1-induced phosphorylation of myosin light chain (pMLC); TGF-β1, however, decreased the ability of a β 2 -agonist to abrogate Cch-induced pMLC [11]. Cch-induced pMLC was inhibited by ISO, forskolin (an adenylyl cyclase activator) and CTX to comparable levels, as shown in Fig. 2. Interestingly, TGF-β1-induced pMLC was also decreased with CTX, forskolin, ISO and SB-431542, a TGF-β1 receptor antagonist. However, ISO and CTX-induced inhibition was less effective as compared with that of SB-431542 or forskolin in blocking TGF-β-induced pMLC (Fig. 2, upper right). Phosphorylation of SMAD3 (pSMAD3) induced by TGF-β1 confirmed engagement of TGF-β1 receptors and activation of downstream signaling pathways (Fig. 2, lower panel). Collectively, these data suggest that TGF-β1 inhibits the ability of β2AR or G αs Fig. 1 Cholera toxin (CTX) increases cAMP levels in HASM, which is blunted by overnight treatment with TGF-β1. A time course of CTX-, a G αs activator, induced cAMP production in HASM (0.25 μg/ml, 0-75 min) was performed in the presence and absence of TGF-β1 (10 ng/ml, 18 h). Data is expressed fold change over vehicle control as mean ± SEM for n = 3 separate cell lines, three additional donors were added to the 45 min to confirm appropriate time point. *p < 0.05 as assessed by one-way ANOVA, and comparisons between two conditions assessed by Student's t-test  HASM were treated with TGF-β1 (10 ng/ml, 18 h), total RNA was isolated, and gene expression was assessed by TaqMan qPCR. Expression of grk2 and grk3 was normalized to endogenous β-actin. Data is representative of n = 5-6 different donors as mean ± SEM, *p < 0.05 by Student's t-test activation, but not forskolin, to increase cAMP levels and diminish pMLC.

TGF-β1 attenuates grk2 and grk3 expression in HASM
One mechanism by which inflammatory mediators can attenuate signaling downstream of the β 2 AR is through phosphorylation of the receptor [20][21][22][23]. Evidence suggests that GRK2 and 3 are associated with the β 2 AR, mediating desensitization through phosphorylation of intracellular portions of the receptor [24,25]. We next examined the effect of TGF-β1 on expression of GRK2 and 3. As shown in Fig. 3, we show that TGF-β1 attenuates, rather than augments, grk2 and grk3 expression. Despite our previous findings showing that β 2 AR phosphorylation is increased following TGF-β1 exposure [26], we determined that the increase in TGF-β1induced β 2 AR phosphorylation is not due to increased expression of GRK2/3.

Dexamethasone rescues TGF-β1-induced decreases in cAMP levels induced by ISO or CTX
Glucocorticoids remain a cornerstone in the management of asthma by decreasing airway inflammation and reversing β2AR desensitization [27,28]. To address whether the steroids also modulate TGF-β1 effects on cAMP levels, HASM cells were pretreated with dexamethasone (DEX) in the presence and absence of TGF-β1 and then treated with either ISO or CTX. As shown in Fig. 4a, TGF-β1 treatment attenuated ISO-induced cAMP by 47.3% and DEX pretreatment significantly rescued TGF-β1's effect by 35.7%. Similarly, TGF-β1 treatment attenuated CTX-induced cAMP by 61.7%, and DEX pretreatment significantly reversed TGF-β1's effect by 43.8% (Fig. 4b). These data demonstrate partial steroid responsiveness of TGF-β1-mediated hyporesponsiveness to β2AR-induced cAMP levels in HASM.
Dexamethasone attenuates TGF-β1-induced pde4d expression in HASM Previously, we determined that the blunted cAMP response to ISO induced by TGF-β1 was dependent upon increased pde4d expression [11]. Given that DEX rescued TGF-β1-induced attenuation of ISO-and CTX-induced cAMP production, we posited that DEX pretreatment would attenuate TGF-β1-mediated pde4d expression. In a dose-dependent manner, DEX pretreatment significantly attenuated TGF-β1-induced pde4d expression in HASM as shown in Fig. 5.

Discussion
Insensitivity to current therapeutics occurs in some patients with severe asthma. Evidence suggests that β 2 AR dysfunction can manifest in an inflammatory milieu due to inflammatory cytokines associated with severe asthma and airway remodeling such as TGF-β1, IL-13, and TNF-α [11,19,29]. Additionally, TGF-β1 stimulation has been shown to attenuate β adrenergic receptor signaling in other cell types [30,31] in addition to HASM. We examined whether generation of cAMP levels in HASM cells after CTX-treatment occurred in the presence and absence of TGF-β1 (Fig. 1b). As we had previously described, β 2 -agonist-induced responses were blunted following TGF-β1 treatment. To characterize mechanisms by which TGF-β1 diminishes β2AR responses, the effects of TGF-β1 on G αs -induced reversal . Data is representative of (a) n = 6-13 (b) or n = 4-8 separate cell lines as mean ± SEM, *p < 0.05 comparing control/ISO/CTX to TGF-β1/ISO/CTX, and TGF-β1/ISO/CTX to TGF-β1/ISO/CTX ± DEX using a one-way ANOVA, and comparisons between two conditions assessed by Student's t-test of agonist-induced HASM cell contractile signaling was examined. We extended our previous findings that phosphorylation of MLC, an important element of agonistinduced HASM contraction, induced by Cch or TGF-β1 was reversed by activation of the β2AR and G αs (Fig. 2). Given these data and our previously published results, these data suggest that the effects of TGF-β1 are more prominent at the receptor and G αs level rather than downstream at the level of adenylyl cyclase, as evidenced by the lack of effect of TGF-β1 on forskolin-induced increases in cAMP levels.
Attenuation of β 2 AR signaling occurs after phosphorylation of the receptor, promoting internalization and desensitization of the β 2 AR [32][33][34]. Activation of GRK2 and GRK3, members of the GRK (G proteincoupled receptor kinase) family, provide a mechanism by which β 2 AR desensitization occurs. We previously demonstrated that TGF-β1 stimulation induces β 2 AR phosphorylation that is consistent with desensitization of the receptor [26]. To assess a potential role for GRKs in attenuating β 2 -agonist-induced relaxation of HASM, grk2 and grk3 expression were assessed with overnight TGF-β1 treatment. In Fig. 3, we demonstrate that TGF-β1 treatment decreased grk2 and grk3 expression. Given these data, it is highly unlikely that upregulation of GRK2/3 expression contributes the hyporesponsiveness to bronchodilators induced by TGF-β1. Current evidence and our previous studies suggest that TGF-β1 likely modulates activity of these types of kinases, rather than modulating GRK expression to induce hyporesponsiveness to β 2 -agonists.
Currently, glucocorticoids like dexamethasone are used to attenuate inflammation associated with asthma as well as restore β 2 AR responsiveness [3,19,35], and reverse the effects of inflammatory mediator-induced β 2 AR dysfunction. In osteoblasts, it has been demonstrated that dexamethasone suppresses the transcriptional activity of, but not expression, Smad3, attenuating TGF-β1-induced alkaline phosphatase activity and type I collagen expression [16]. Additionally, dexamethasone also repressed transcriptional activation of PAI-1 through inhibition of Smad3/4 by direct interaction between the glucocorticoid receptor and Smad3 [17]. We previously showed that TGF-β1-induced attenuation of ISO reversal of Cchinduced pMLC was Smad3-dependent in HASM. We also showed that TGF-β1 treatment increased expression of pde4d, suggesting a role for phosphodiesterases in TGF-β1-induced hyporesponsiveness to bronchodilators. Given our data and others' observations concerning the effects of glucocorticoids on TGF-β1/Smad3-dependent signaling, we posited that dexamethasone would thereby reverse TGF-β1-and Smad3-induced attenuation of ISO-induced cAMP accumulation and pde4d expression in HASM. Figure 6 depicts a model of mechanisms underlying glucocorticoid-mediated rescue of TGF-β1-induced β 2 AR hyporesponsiveness. We demonstrate that TGF-β1induced attenuation of β 2 -agonist-and G αs -induced cAMP accumulation in HASM can be rescued by treatment with dexamethasone ( Fig. 4a & b). Our data also show that DEX treatment reversed TGF-β1-induced pde4d expression (Fig. 5). Consistent with these data, we previously demonstrated that DEX (1 uM, 18 h) stimulation alone has little effect pde4d expression [36]. We recognize that changes in transcript expression do not necessarily translate to changes in protein expression or PDE4D activity. Previous work in HASM has demonstrated that PDE4D5 is expressed and, in part, controls cAMP production following β 2 agonist stimulation [37,38]. As in our study, Billington et al. demonstrated transcript expression, but not protein expression or enzymatic activity, of the specific PDE isoform. Interestingly, Niimi et al. demonstrated that transcript expression of PDE4D isoforms in HASM was altered, but not protein or enzymatic activity [39]. Extensive work by the Houslay group has shown the difficulty of isolating specific isoforms and splice variants of PDE enzymes, as well as the difficulty of enzymatic activity assessment [40]. Given the findings of Billington, Niimi, and Trian, and the complexities of the assays performed by the Houslay group, we recognize the limitations to our work, but feel that our results likely translate to functional outcomes affecting bronchomotor tone in the context of TGF-β1-induced hyposensitivity to β 2 agonists and cAMP mobilizers.
Despite the fact that we showed that TGF-β1 attenuates CTX-induced cAMP production, and that DEX Data is represented as % of pde4d expression induced by TGF-β1, with pde4d expression normalized to β-actin. Data is representative of n = 4-6 separate cell lines as mean ± SEM, *p < 0.05 using a one-way ANOVA, and comparisons between two conditions assessed by Student's t-test. TGF-β1 vs TGF-β1+ DEX (30 nM), p = 0.08 rescues TGF-β1-mediated attenuation of both ISO-and CTX-induced cAMP production, it remains to be seen if these mechanisms are operative in vivo. Despite this limitation, we have shown effects of TGF-β1 on HASM to be recapitulated in small airways derived from human lungs [10]. Additionally, while it would be interesting to study this phenomenon in asthma-derived HASM, we and others have demonstrated that β 2 -agonist-induced cAMP production in these cells is already blunted due partially to increased PDE expression [26,38]. Therefore, exposure of asthma-derived HASM to TGF-β1 will likely have little effect on modulating β 2 -agonist-induced cAMP production.

Conclusion
Regardless of evidence that steroids may not reverse the TGF-β1-induced remodeling effects [41] in asthma, our findings suggest that in asthma patients with high levels of TGF-β1, steroids may be an effective treatment to reverse β 2 AR hyporesponsiveness observed in these patients.
Additional file 1: Table S1. Donor demographics for cAMP and pde4d expression studies. All cells were derived from subjects with no history of chronic disease.  We previously demonstrated that TGF-β1 induces Smad2/3 activation to increase pde4d expression, leading to increased cAMP hydrolysis and attenuating HASM cell relaxation responses. We posited that DEX binds to the glucocorticoid receptor (GR), inhibiting both increased pde4d expression and rescuing TGF-β1-induced attenuation of β 2 AR and G αs -induced cAMP production. AC = Adenylyl Cyclase; β 2 AR = β 2 -adrenergic receptor; DEX = Dexamethasone; G αs = Stimulatory G α protein; GR = Glucocorticoid Receptor; PDE4D = Phosphodiesterase 4D; TBR-I/II = TGF-β receptor I/II; pMLC = phosphorylated myosin light chain