Functional K_Ca3.1 channels regulate steroid insensitivity in bronchial smooth muscle cells

Abstract

Identifying the factors responsible for relative glucocorticosteroid (GC) resistance present in patients with severe asthma and finding tools to reverse it are of paramount importance. In asthma there is in vivo evidence of GC-resistant pathways in airway smooth muscle (ASM) bundles which can be modelled in vitro by exposing cultured ASM cells to TNFα/IFNγ. This drives GC insensitivity via protein phosphatase-5 (PP5)-dependent impairment of GC receptor (GR) phosphorylation. Here, we investigated whether K_Ca3.1 ion channels modulate the activity of GC-resistant pathways using our ASM model of GC insensitivity. Immunohistochemical staining of endobronchial biopsies revealed that K_Ca3.1 channels are localized to the plasma membrane and nucleus of ASM in both healthy controls and asthmatic patients, irrespective of disease severity. Western blot assays and immunofluorescence staining confirmed the nuclear localisation of K_Ca3.1 channels in ASM cells. The functional importance of K_Ca3.1 channels in the regulation of GC-resistant chemokines induced by TNFα/IFNγ was assessed using complementary inhibitory strategies including K_Ca3.1 blockers (TRAM-34 and ICA-17043) or K_Ca3.1-specific shRNA delivered by adenoviruses. K_Ca3.1 channel blockade led to a significant reduction of fluticasone-resistant CX3CL1, CCL5 and CCL11 gene and protein expression. K_Ca3.1 channel blockade also restored fluticasone-induced GRα phosphorylation at ser211 and transactivation properties via the suppression of cytokine-induced PP5 expression. The effect of K_Ca3.1 blockade was evident in ASM cells from both healthy controls and asthmatic subjects. In summary K_Ca3.1 channels contribute to the regulation of GC-resistant inflammatory pathways in ASM cells: blocking K_Ca3.1 channels may enhance corticosteroid activity in severe asthma.

INTRODUCTION

Although many patients with asthma are well controlled with inhaled glucocorticoid (GC) therapy, approximately 5 to 10% of patients have difficult-to-control or severe disease that is poorly responsive to this treatment. These patients are responsible for a disproportionate share of asthma-related health care costs and morbidity (1). This represents a significant unmet clinical need and novel therapies are urgently required (1).

Despite significant progress in the field, the precise molecular mechanisms mediating GC insensitivity in severe asthma are poorly understood. Several potential mechanisms have been postulated, identified mostly from studies in immune cells including alveolar macrophages and circulating peripheral blood mononuclear cells (PBMCs) obtained from patients with GC insensitivity (2–6). In addition to infiltrating inflammatory cells, the airway smooth muscle (ASM) is another key player in severe asthma pathophysiology, demonstrating heightened sensitivity to both direct and indirect contractile stimuli leading to exaggerated airway narrowing and airflow obstruction (7). The therapeutic benefit in severe asthmatics provided by bronchial thermoplasty, a therapy that attenuates bronchoconstriction via reduction of ASM mass, has strongly supported the notion that ASM is playing a central role in the pathogenesis of the severe disease (8–11). In addition to its central role in bronchoconstriction, the role of ASM in the pathogenesis of severe asthma could derive also from its immunomodulatory function as it secretes a variety of pro-inflammatory mediators (12). This is evident both in vitro using cultured ASM cells and in vivo in the ASM bundles of asthmatic patients (12).

Whether the pathogenesis of severe asthma is driven by the steroid-resistant production of proteins from ASM cells represents an interesting hypothesis. Indeed, previous reports convincingly showed that despite patients taking high doses of inhaled or oral GC, there is ongoing expression of different “pro-asthmatic” proteins in asthmatic ASM bundles, including the chemokines CX3CL1 (13), CCL11 (14), CCL15 (15), and CCL19 (16), and ADAM33 and ADAM8 (17, 18). These studies provide indisputable in vivo evidence for the existence of steroid-resistant pathways in ASM that are potentially driving inflammatory processes and ASM contractile dysfunction in asthmatic airways. A better understanding of the underlying mechanisms driving these steroid-resistant pathways in ASM could therefore lead to novel therapeutic approaches.

The intermediate conductance Ca²⁺-activated K⁺ (K_Ca) channel K_Ca3.1 channel (also known as IK1, SK4 or KCNN4) is closely associated with the progression of number of human diseases, and is expressed by human ASM cells (19, 20). In human ASM cells, a K_Ca3.1 blocker attenuated mitogen-induced ASM cell proliferation (20). Inhibitors of K_Ca3.1 channels such as TRAM-34 are effective in treating various inflammatory diseases in preclinical models including atherosclerosis (21), neurodegenerative disorders (22), autoimmune encephalomyelitis (23), and coronary vasculoproliferative diseases (24). With respect to asthma, we recently provided the first evidence that TRAM-34 prevents the development of eosinophilic inflammation, airway hyperresponsiveness and airway remodeling in a murine model of allergic asthma (25). The underlying mechanisms by which K_Ca3.1 channels contribute to the pathogenesis of allergic asthma are yet to be defined but we have shown that K_Ca3.1 channels regulate mast cell degranulation and migration (26–29), and fibrocyte migration (30). Others have implicated K_Ca3.1 channels in the migration of lung dendritic cells to CCL19 and CCL21 (31). These observations demonstrate that activation of K_Ca3.1 channels on structural (smooth muscle/fibroblasts) airway cells may represent an important pathway driving key features of allergic asthma.

In the present study, we made the surprising finding that K_Ca3.1 channels are not only essential in driving the production of GC-resistant chemokines by ASM cells but also contribute to the reduced ASM sensitivity to GC therapy via the upregulation of serine/threonine phosphatase PP5. This is the first report to demonstrate a functional interaction between K_Ca3.1 channels and the impairment of GC function. This study uncovers a novel molecular mechanism contributing to the development of GC insensitivity, a defining feature of severe asthma.

MATERIALS AND METHODS

Human subjects

Asthmatic subjects, COPD patients and healthy volunteers were recruited. Asthmatic and COPD subjects had a consistent history and objective evidence of asthma or COPD as described previously (32). Asthma severity was defined by British Guideline on the Management of Asthma treatment steps: mild = step 1, β₂-agonist only; moderate = steps 2 and 3, inhaled corticosteroid ≤800 μg beclomethasone equivalent per day ± long-acting β₂-agonist; severe = steps 4 and 5 (33). Five out of 7 patients at step 4/5 undergoing immunohistological staining met the American Thoracic Society criteria for refractory asthma (1). All COPD patients used for the study were classified as GOLD I and II. The studies were approved by the Leicestershire, Northamptonshire, & Rutland Research Ethics Committee (references: 4977, 04/Q2502/74 and 08/H0406/189). Written informed consent was gained from all participants prior to their involvement. The demographics of the patients taking part are shown in Tables 1 and 2.

Demographics of the patients used for in the immunohistochemistry studies

	Healthy control subjects	Patients with mild-moderate asthma (BTS steps 1–3)	Patients with severe asthma (BTS steps 4 and 5)	p value
Number	7	5	7	-
Age (mean ± SEM)	35.71 ± 6.98	36 ± 7.86	42.29 ± 4.18	0.590
Sex (M/F)	2/5	2/3	3/4	0.843
Asthma duration (mean ± SEM)	N/A	17 ± 5.38	33.86 ± 5.63	0.064
Inhaled corticosteroid dose (beclomethasone equivalents†) (mg)	N/A	512.5 ± 234.9	1400 ± 130.9	0.006
Number at BTS step 5	N/A	0	4	-
Number on long-acting beta-agonist	N/A	2	7	-
Exacerbations in last year (median [range])	N/A	0 [0–1]	1 [0.0–3.0]	0.238
Mean daytime symptom score (median [range])	N/A	0.25 [0.18–0.54]	0.93 [0.14–1.39]	0.176
Mean daily night time symptom score (median [range])	N/A	0 [0.0–0.0]	0.28 [0.0–1.07]	0.109
Reliever use/week (median [range])	N/A	1.5 [0.0–5.25]	22.0 [6.0 –37.0]	0.037
Sputum eosinophil count (%) (geometric mean [95% CI])	0.34 [0.20–0.56]	8.86 [2.08–37.75]	2.07 [0.43–9.98]	0.006
PEF amplitude % of the mean (mean ±SEM)	N/A	20.08 ± 6.43	34.50 ± 9.45	0.272
FEV1 (% predicted)	102.4 ± 3.45	91.80 ± 7.37	66.57 ± 6.71	0.006
FEV1/FVC (%)	84.0 ± 3.90	76.40 ± 5.17	59.57 ± 6.47	0.018
PC20 methacholine (mg/ml)(geometric mean [95% CI])	>16	0.6841 [0.02–30.49]	0.11 [0.02–0.72]	0.123
Serum IgE (kU/L) (geometric mean [95% CI]	31.59 [12.57–79.4]	501.5 [43.07–5840]	203.1 [28.98–1424]	0.021
Number with positive skin prick test	5	4	5	0.368
Number with positive skin prick test to Aspergillus fumigatus	0	0	1	0.368

Demographics of the patients used for the chemokine expression assays/immunoblot and immunostaining studies

	Controls	Asthma	COPD	P Value
Number	12	10	6	-
Gender (M/F)	5/7	5/5	6/0
Age (mean ± SEM)	42.25 ± 4.15	49.4 ± 2.98	66 ± 4.31	0.0046
FEV1 (± SEM)	3.7 ± 0.27	2.48 ± 0.22	1.9 ± 0.31	0.0004
FEV1 % predicted (± SEM)	88.75 ± 4.08	70.92 ± 9.97	61.5 ± 7.04	0.0116
FEV1/FVC (%) (± SEM)	88.7 ± 8.3	64.4 ± 3.32	53.82 ± 5.16	0.0044

Fiberoptic bronchoscopy

Subjects underwent fiberoptic bronchoscopic as described previously (32), and according to British Thoracic Society guidelines (34). Bronchial mucosal biopsy specimens were taken from the right middle lobe and lower lobe carinae, fixed in acetone, and embedded in glycol methacrylate (GMA) (35). Further biopsies were dissected for the isolation and culture of ASM (36).

K_Ca3.1 staining in endobronchial biopsies

Sequential 2μm sections were cut from GMA-embedded bronchial biopsies and immunostained as described previously (35) using a rabbit polyclonal anti-human K_Ca3.1 antibody (M20, 2.7 μg/ml, gift from Dr Chen, GlaxoSmithKline, Stevenage, UK, (37)), and isotype control rabbit IgG (2.7 μg/ml) (R&D Systems). Cells staining positively for K_Ca3.1 in the ASM compartment were quantified using a semi-quantitative intensity score: 0 = no staining, 0.5 = sparse and patchy, 1.0 = diffuse but weak, 2.0 = diffuse. The percentage of nuclei that were stained positive within the ASM cells was also calculated. All sections were counted by a blinded observer.

Culture of airway smooth muscle cells

Primary human ASM cells were obtained from healthy subjects, COPD and asthmatic patients isolated from endobronchial biopsies as previously described (36). Briefly, pure ASM bundles in airways were dissected free of surrounding tissue. The small muscle bundles were cultured in DMEM supplemented with 10% FBS, 4 mM L-glutamine, 100 U/ml penicillin, 100 mg/ml streptomycin, and 0.25 μg/ml amphotericin. ASM characteristics were determined by immunofluorescence and light microscopy with α-smooth muscle actin-FITC direct conjugate and myosin indirectly conjugated with FITC (Sigma-Aldrich). Cells were used from passage 2–6.

Treatment of airway smooth muscle cells

Cells were pre-treated with the specific K_Ca3.1 channel blockers TRAM-34 (200 nM) (a gift from Heike Wulff, UC Davis, CA, USA) (38) and ICA-17043 (60 nM) (a gift from Icagen,Inc., Durham, NC, USA) (39) alone or in combination with fluticasone (FP, 100 nM) for 2 h and then stimulated with TNFα (10 ng/ml)/IFNγ (500 IU/ml) for an additional 24 h. Cell viability was assessed in cells from 2 asthmatics (each performed in triplicate) using the MTT (3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide, a yellow tetrazole) assay as described in our previous report (41).

ELISA

ELISAs for assessing chemokine levels were performed on 100 μl cell supernatants using the R&D System Duoset kits (Minneapolis, MN, USA) according to the manufacturer's instructions (40). DMSO controls were performed, with the final concentration of DMSO 0.1% in all control and drug wells.

qPCR

Total RNA was extracted using RNeasy plus mini kit (Qiagen #74134). cDNAs were synthesised using a first stand cDNA synthesis Kit (Fermentas life science). Real time quantitative PCR was performed using Brilliant SYBR® Green QPCR Master Mix (Agillent) using a Stratagene Mx3000P (Santa Clara, CA, USA). Each sample from different patients was assayed in duplicate. For relative quantification, the C_T method: ratio = 2-[(Ct (gene of interest)-Ct (β actin as internal control)] (2^−ΔΔCt) was used as described in (42) to determine the fold change due to the different treatment conditions (fluticasone alone, K_Ca3.1 blockers alone, and both drugs in combination). The primers were as followed: β-actin, forward CCCAAGGCCAACCGCGAGAAGAT, reverse: GTCCCGGCCAGCCAGGTCCAG, CX3CL1, forward: GCTGAGGAACCCATCCAT, reverse: GAGGCTCTGGTAGGTGAACA and CCL5, forward: TCTGCGCTCCTGCATCTG, reverse: GGGCAATGTAGGCAAAGCA, CCL11 forward: AAT GTC CCC AGA AAG CTG TG, CCL11, reverse: TCC TGC ACC CAC TTC TTC TT. GILZ: Forward 5'-TCTGCTTGGAGGGGATGTGG-3' Reverse 5'- ACTTGTGGGGATTCGGGAGC-3'.

Immunofluorescence

ASM cells grown on the chamber slides were fixed with 10% neutral buffered formalin for 15 min at room temperature, washed 3 times with Phosphate Buffered Saline (PBS), permeabilized using 90% ice cold methanol for 20 min and washed again 3 times with PBS. The cells were then incubated with 3% Bovine Serum Albumin/PBS (BSA/PBS) for 1 h before anti-K_Ca3.1 (2.5 μg/ml, Biorbyt, Cambridge, UK), anti-HDAC-1 (1 μg/ml, Cell Signaling, Danvers, USA) or appropriate isotype-matched control antibodies (all prepared in 1% BSA/PBS) were added overnight at 4°C. After 3 washes, Alexa-Fluor®488 goat anti-rabbit IgG antibody (Invitrogen Ltd, Paisley, UK) was added at 1:300 dilution for 40 min then washed with PBS and mounted using vectamount (Vector Laboratories Ltd, Peterborough, UK). Images of the cells were viewed on a confocal microscope (Leica TCS SP5 confocal) and captured using Leica LAS AF software (Leica, UK).

Western blot

Immunoblots on total cell extracts from ASM cells were performed as described previously (43) using anti-GR-S211 and anti-GR antibodies (Cell signaling, Danvers, MA). To ensure equal loading, the membranes were stripped and reprobed with anti-β-actin antibody (Santa Cruz Biotechnology, Santa Cruz, CA). Immunoblots on nuclear fractions used samples prepared using the extraction reagent kit (NE-PER) (Thermo-Fisher Scientific). 30 μg of nuclear extracts was analyzed by 10% SDS-PAGE and assayed by immunoblot using anti-K_Ca3.1 antibody from Sigma-Aldrich (1:500, Dorset, England) and anti-β-actin (1:200 dilution). The anti-K_Ca3.1 antibody from Biorbyt, Cambridge, UK (1:500 dilution) was used for assessing K_Ca3.1 expression in total cell lysates (60 μg) in the silencing studies. Anti-mouse IgG-HRP or anti-rabbit IgG-HRP (1:5000 dilution, Santa Cruz Biotech.) were used as secondary antibodies. The bands were visualized by the enhanced chemiluminescence system (Amersham Biosciences) and autoradiographed.

shRNA gene knockdown of K_Ca3.1 channels

K_Ca3.1 (V1 and V2), control (V5) and GFP-shRNA-expressing adenoviruses (Ad5C20Att01) were purchased from BioFocus DPI (Leden, The Netherlands). Preliminary transduction experiments using this adenovirus expressing cDNA for GFP demonstrated high transduction efficiency (~80% of cells). Gene knockdown of K_Ca3.1 channels using different shRNA adenoviruses was confirmed using 2×10⁵ ASM cells transduced using a multiplicity of infection of 30 (30 MOI) and subsequent Western blotting. For experiments, 24 h after transduction, the cells were serum-deprived for 3 h and then exposed to fluticasone (100 nM) for 2 h followed by addition of control media or TNFα/IFNγ for an additional 22 h. After 24 h, proteins supernatants were collected for ELISA assays (CCL5, CX3CL1, CCL11) while cells were lysed for total cell extract preparation as described above.

Flow cytometry

ASM cells were fixed and permeabilized in 4% paraformaldehyde/0.1% saponin for 15 minutes on ice. Cells were then stained with either 2 μg/ml rabbit anti-human PP5 antibody (Cell Signaling, Danvers, MA USA) or isotype matched control (rabbit IgG-Immunostep, Salamanca Spain) overnight, followed by secondary sheep anti-rabbit IgG:RPE antibody (Cell Signalling, Danvers, MA USA) for 2 hr. Staining was examined by flow cytometry using the Becton Dickinson FACScan (Oxford, UK).

Patch clamp electrophysiology

The whole-cell variant of the patch-clamp technique was used as previously (20). Briefly, ASM cells were cultured until confluent and then maintained in ITS media as normal. Cells were incubated overnight with TNFα and IFNγ for 24hr with control cells being maintained under identical conditions without added cytokines. Prior to electrophysiological recording the cells were trypsinised and re-suspended in ASM culture media. Cells were pipetted into a heated (30°C) recording chamber ready for patching. Individual (non-adhered) cells were voltage-clamped using the whole-cell variant of the patch-clamp technique. Briefly voltage commands were applied to potentials ranging from −100 to +100mV from a holding potential of −20mV. Membrane currents were recorded and subsequently plotted against command potential. Solutions used: External (mM) NaCl (140), KCl (5), MgCl2 (1), CaCl2 (2); pH was adjusted to 7.4 using NaOH. Internal (mM) KCl (140), MgCL2 (2), HEPES (10), ATP (2), GTP (0.1); pH was adjusted to 7.3 using KOH.

Statistical analysis

For cell culture studies across group comparisons were made using one-way ANOVA or repeated measures ANOVA as appropriate, with the Bonferroni post hoc test for comparisons between specific groups. Immunohistochemical staining was analysed across groups using the Kruskall Wallis test. p<0.05 was taken as statistically significant.

RESULTS

Expression of K_Ca3.1 in ASM bundles

The clinical characteristics of the subjects assessed by immunohistochemistry are described in Table 2. There was no immunostaining in isotype control sections (Fig.1A), but K_Ca3.1 immunoreactivity was positive in ASM bundles (Fig.1B–D), irrespective of disease status (Fig.1E). The overall intensity of staining of K_Ca3.1 in the ASM was not different between healthy subjects (Fig.1B) and asthmatic patients with moderate (Fig.1C) or severe disease (Fig.1D)(quantification not shown). Interestingly, not only was K_Ca3.1 localised to the plasma membrane and cytoplasm (consistent with channel trafficking), but it was also localised to the nuclear membrane (Fig.1E). The immunostaining taken at a higher magnification clearly shows a marked expression of K_Ca3.1 channel in nucleus of three different cells in one asthmatic patient (X1000). The proportion of ASM cells expressing the K_Ca3.1 channel in the nucleus was not different between normal and asthmatic conditions (Fig.1F).

Figure 1. In vivo K_Ca3.1 expression in human ASM bundles.

Representative immunostaining in bronchial biopsies for (A, X200) isotype control antibody, K_Ca3.1 in a healthy control (B, X400), in an asthmatic treated with ICS (C, X400), or with ICS + LABA (D, X400). (E) Representative immunostaining at higher magnification (X1000) showing the nuclear expression of K_Ca3.1 channel in vivo in 3 different muscle areas. The upper left image is the tissue incubated with the isotype-matched control antibody. (F) The percentage of ASM cell nuclei staining for K_Ca3.1 in normal subjects (n=7), and moderate (n=5) and severe asthma (n=6) patients. Two sections at least 10 microns apart were analysed for each subject, and the mean of the two sections taken as the value for each subject. A minimum of 15 cells per section were counted (range 15–100 cells, median 49.5 cells). Data are expressed as means ± S.E.M. Ep, epithelium, ASM, Airway smooth muscle.

Nuclear expression of K_Ca3.1 in cultured ASM cells

To further investigate whether K_Ca3.1 channel was also present in the nuclear compartment in cultured ASM cells, we conducted immunofluorescence staining in permeabilized cells derived from 3 patients. As shown in Fig.2A K_Ca3.1 channel was identified within the nucleus in ASM cells. Staining for HDAC1 (Fig.2B), a protein exclusively found in the nucleus (44), was used as a positive control for the immunostaining procedure. The intracellular expression of K_Ca3.1 in ASM cells is highly novel although one report identified K_Ca3.1 channels in mitochondrial membranes in human colon carcinoma 116 cells (45). We also performed western blots directly on isolated nuclear extracts to provide additional evidence for the nuclear expression of K_Ca3.1 channel in human ASM cells. Interestingly, the nuclear fraction contained discreet bands recognized by an anti-K_Ca3.1 antibody of 47 and 53 kDa, as described previously in the cytoplasmic extracts of several cell types (provide several (20, 37, 46). (Fig. 2C). The predicted size of K_Ca3.1 is 48 kDa. The larger band is likely to represent alternative glycosylation, although several smaller splice variants are also described (46). The quantitative analysis of K_Ca3.1 expression in the nuclear fractions by immunoblot assay revealed no significant changes between healthy and asthmatic subjects (Fig.2C, bottom panel). This confirms the nuclear expression of intracellular K_Ca3.1 in human ASM cells.

Figure 2. Nuclear expression of K_Ca3.1 channel in cultured human ASM cells.

Permeabilized ASM cells were stained for K_Ca3.1 (A), the nuclear marker HDAC1 (B) or incubated with the corresponding isotype-matched antibodies used as negative controls. DAPI counterstaining was performed on the same cell population to demonstrate the nuclear location of K_Ca3.1. Images are representative of experiments performed on cells from 3 different healthy subjects (C) (Top) K_Ca3.1 expression in the nucleus was also confirmed by assaying ASM nuclear fractions for K_Ca3.1 by immunoblot analysis. (Bottom) Scanning densitometric measurement of K_Ca3.1 expression normalized over the corresponding β-actin showed no significant difference in K_Ca3.1 expression in ASM nuclear fractions between non-asthmatic and asthmatic subjects. Results are representative of blots performed in 6 subjects (3 healthy and 3 asthmatics). (D), Current voltage curves showing the lack of detectable K_Ca3.1 currents in ASM cells at baseline and after cytokine treatment. Data are expressed as means ± S.E.M of measurement performed in n=16 cells per condition taken from a healthy subject.

Using patch clamp electrophysiology, we also assessed whether channel activity was affected in cells treated with TNFα/IFNγ. Fig. 2D compares the average current-voltage (I/V) curves performed in 16 healthy cells from diluent or cytokine-treated conditions for 24 h. In agreement with our previous report (20), we found no evidence of K_Ca3.1 currents in non-stimulated cells. Interestingly, K_Ca3.1 currents were also not noticeable in cells treated with TNFα/IFNγ, suggesting that cytokines do not stimulate plasma membrane K_Ca3.1 channel activity in ASM cells.

Production of steroid-resistant chemokines induced by TNFα/IFNγ in ASM cells from healthy controls and patients with asthma and COPD

We first determined whether production of 4 different chemokines, CCL5, CCL11, CX3CL1 and CXCL10 in response to TNFα/IFNγ was affected by disease status (Fig. 3). We found that levels of CCL5 (Fig.3A), CCL11 (Fig.3B), and CXCL10 (Fig.3C) produced in response to TNFα/IFNγ at 24 hr were similar in the different groups studied. Thus net increases (induced minus basal levels) of CCL5 were 7269 ± 1841, 5968 ± 1388, 7280 ± 2297 pg/ml in healthy, asthma and COPD subjects, respectively (Fig.3A). Net increases of CCL11 were 417.2 ± 93, 266.6 ± 92, 317 ± 121 pg/ml in healthy, asthma and COPD subjects, respectively (Fig.3B). Net increases of CXCL10 were 72666 ± 9740, 71478 ± 11260, 79401 ± 11272 pg/ml in healthy, asthma and COPD subjects, respectively (Fig.3C). Interestingly, we found that ASM cells from COPD patients produced significantly less CX3CL1 in comparison to healthy controls (774 ± 142 versus 4763 ± 1400 pg/ml, p<0.05) while no difference was noticed when compared to levels produced by asthmatics (2927 ± 501 pg/ml) (Fig.3D).

Figure 3. Chemokine production induced by TNFα/IFNγ in ASM cells from health, asthma and COPD subjects.

Cells were stimulated with TNFα/IFNγ for 24 h. CCL5, CX3CL1, CCL11 and CXCL10 levels in the supernatants were assessed by ELISA assays as described in the methods section. Data are expressed as means ± S.E.M of the net TNFα/IFNγ-induced increase of CCL5 (A), CCL11 (B), CX3CL1 (C), CXCL10 (D) in n=5 controls, 7 asthmatics and 5 COPD patients, all performed in triplicate. All chemokines were significantly increased compared to unstimulated control. NS represents non-significance between groups, ***P<0.01 compared to healthy cells.

K_Ca3.1 channel inhibition differentially suppresses the production of TNFα/IFN-induced steroid-resistant chemokines

We next investigated the role of K_Ca3.1 inhibitors in our ASM cell model of GC resistance (TNFα/IFNγ-treated cells) by assessing the expression of 4 different chemokines in ASM cells from healthy (n=4), asthmatic (n=6) and COPD subjects (n=4). As reported in tracheal ASM cells (12), induction of CX3CL1, CCL5, CCL11 and CXCL10 by TNFα/IFNγ were completely resistant to fluticasone treatment in ASM cells derived from the bronchial tree (Fig.4–7). Chemokine production was not affected in cells treated with 0.1% DMSO, the final concentration of the inhibitor solvent.

Figure 4. Inhibition of GC-resistant CCL5 by K_Ca3.1 blockers.

Cells pre-treated with TRAM-34 (200 nM) or ICA-17043 (60 nM) with or without fluticasone (100 nM, 2 h) were stimulated with TNFα/IFNγ for 24 h. CCL5 levels in the supernatants were assessed by ELISA as described in the methods section. Data are expressed as means ± S.E.M % of TNFα/IFNγ-induced CCL5 levels. Healthy controls (A, n=4), asthmatics (B, n=6) and COPD patients (C, n=4). *P < 0.05, **P<0.01 compared to TNFα/IFNγ/DMSO control.

Figure 7. Lack of inhibition of GC-resistant CXCL10 by K_Ca3.1 blockers.

Cells pre-treated with TRAM-34 (200 nM) or ICA-17043 (60 nM) with or without fluticasone (100 nM, 2 h) were stimulated with TNFα/IFNγ for 24 h. CXCL10 levels in the supernatants were assessed by ELISA as described in the methods section. Data are expressed as mean ± S.E.M % of TNFα/IFNγ-induced CXCL10 levels. Healthy controls (A, n=4), asthmatics (B, n=6) and COPD patients (C, n=4). None of the values were significantly different when compared to TNFα/IFNγ/DMSO control.

The specific K_Ca3.1 blockers TRAM-34 (200 nM) (38) or ICA-17043 (60 nM) (39) alone did not affect TNFα/IFNγ-induced production of CCL5 (Fig.4) or CCL11 (Fig.5) with the exception of cells derived from asthmatic patients (Fig.5B). The inhibitory effect was further enhanced by more than 90% in the presence of fluticasone (Fig.5B). TNFα/IFNγ–induced production of CCL5 in all tested individuals (Fig. 4) or CCL11 in healthy and COPD subjects (Fig.5A–C) was only inhibited when fluticasone was combined with either K_Ca3.1 channel blocker. Fluticasone and ICA-17043 in combination reduced TNFα/IFNγ-induced CCL5 levels to 57 ± 12%, 61 ± 14.5%, and 52 ± 14% in ASM cells from healthy, asthma and COPD subjects, respectively (Fig.4). Fluticasone and TRAM-34 reduced TNFα/IFNγ-induced CCL5 levels to 57 ± 11%, 51 ± 21%, and 54 ± 9.6% in cells from healthy, asthma and COPD subjects, respectively (Fig.4). Production of CCL11 by TNFα/IFNγ was reduced to 36 ± 14%, 12 ± 14%, and 28 ± 12% of control by the fluticasone and ICA-17043 combination in cells from healthy, asthma and COPD subjects, respectively. TNFα/IFNγ-induced CCL11 production was reduced to 38.5 ± 14.2%, 9 ± 9%, and 31 ± 15% by the fluticasone and TRAM-34 combination in cells from healthy, asthma and COPD subjects, respectively (Fig.5).

Figure 5. Inhibition of GC-resistant CCL11 by K_Ca3.1 blockers.

Cells pre-treated with TRAM-34 (200 nM) or ICA-17043 (60 nM) with or without fluticasone (100 nM, 2 h) were stimulated with TNFα/IFNγ for 24 h. CCL11 levels in the supernatants were assessed by ELISA as described in the methods section. Data are expressed as mean ± S.E.M % of TNFα/IFNγ-induced CCL11 levels. Healthy controls (A, n=4), asthmatics (B, n=6) and COPD patients (C, n=4). *P < 0.05, ***P<0.001 compared to TNFα/IFNγ/DMSO control.

Interestingly, induction of CX3CL1 by TNFα/IFNγ in cells from healthy, COPD and asthma subjects was reduced to 31.4 ± 12.1%, 8.8 ± 3.4%, and 33.2 ± 8.9% by ICA-17043 alone (60 nM) and to 28.9 ± 10.8%, 21.1 ± 10.9%, and 26.4 ± 14.7% by TRAM-34 alone (200 nM) (Fig.6A–C). The degree of CX3CL1 inhibition by K_Ca3.1 blockade was not changed by the presence of fluticasone (Fig.6A–C, 2 last columns).

Figure 6. Inhibition of GC-resistant CX3CL1 by K_Ca3.1 blockers.

Cells pre-treated with TRAM-34 (200 nM) or ICA-17043 (60 nM) with or without fluticasone (100 nM, 2 h) were stimulated with TNFα/IFNγ for 24 h. CX3CL1 levels in the supernatants were assessed by ELISA as described in the methods section. Data are expressed as mean ± S.E.M % of TNFα/IFNγ-induced CX3CL1 levels. Healthy controls (A, n=4), asthmatics (B, n=6) and COPD patients (C, n=4). *P < 0.05, **P<0.01, ***P<0.01 compared to TNFα/IFNγ/DMSO control.

In contrast, production of CXCL10 by TNFα/IFNγ was not affected by either K_Ca3.1 inhibitors used alone or in combination with fluticasone (Fig. 7), ruling out a toxic or non-specific effect of K_Ca3.1 inhibitors. Measuring cell viability using the MTT assay did not show any signs of cytotoxicity in cells treated with the above concentrations of K_Ca3.1 inhibitors. No differences were observed between relative MTT-reducing activities in control cells (with 0.1% DMSO) and in cells treated with either ICA-17043 or TRAM-34 that were 0.344 ± 0.019, 0.317 ± 0.022, and 0.346 ± 0.063, respectively.

shRNA silencing of K_Ca3.1 channels

To confirm the previous findings with pharmacological blockers of K_Ca3.1, we also used knockdown of K_Ca3.1 channels using shRNA delivered by adenoviruses. A shRNA control adenovirus (V5) did not affect K_Ca3.1 channel protein expression (Fig. 8A), but this was completely decreased by K_Ca3.1 shRNAs V1 and V2 (Fig. 8A, top panel). None of the shRNAs affected β-actin expression (Fig. 8A). TNFα/IFNγ treatment alone or in the presence of fluticasone had no effect on K_Ca3.1 expression or on the efficacy of shRNA adenoviruses to silence the channel (Fig. 8A, bottom panel). The following functional assays were then performed using shRNA-K_Ca3.1 adenoviruses V1 and V2, and shRNA-control adenovirus (V5) as a control.

Figure 8. shRNA K_Ca3.1 modulates TNFα/IFNγ-induced chemokine expression.

A, (Top) Representative immunoblot of K_Ca3.1 protein expression in cells transduced with K_Ca3.1 shRNA (V1 and V2) and control (V5) adenoviruses in the presence or absence of cytokines and fluticasone (FP). (Bottom), scanning densitometric measurement of K_Ca3.1 expression by immunoblot assays normalized over the corresponding β-actin showed complete knockdown of K_Ca3.1 expression in ASM cells using shK_Ca3.1 V1 and V2 when compared to control adenovirus. Note that the TNFα/IFNγ combination had no effect on K_Ca3.1 expression. Results are shown as means ± S.E.M of blots performed in 3 healthy donors. B–C, Effect of the same adenoviruses on TNFα/IFNγ-induced expression of CX3CL1 (B) and CCL5 (C) assessed by ELISA. Results are shown as means ± S.E.M of experiments performed in triplicate in 3 donors. **P< 0.01 compared to respective control shRNA.

K_Ca3.1 downregulation in ASM cells attenuates TNFα/IFNγ-induced steroid-resistant chemokine expression

As in untransduced ASM cells, expression of CX3CL1 (Fig. 8B) and CCL5 (Fig. 8C) in V5 control transduced cells was increased by TNFα/IFNγ and unaffected by fluticasone. In contrast, CCL5 production was inhibited by approximately 30% by K_Ca3.1 shRNAs V1 and V2, with further inhibition of secretion with the addition of fluticasone to almost 50% (Fig. 8C). In agreement with the data obtained with the soluble inhibitors, CX3CL1 expression was inhibited by K_Ca3.1 shRNAs V1 and V2 irrespective of fluticasone treatment (Fig. 8B). The % inhibition of CX3CL1 induced by shRNAs V1 in the absence of the presence fluticasone when normalized to CX3CL1 production in the presence of the control shRNA was 68% and 49%, respectively. The % inhibition of CX3CL1 induced by shRNAs V2 was 45% and 30% in the absence of the presence of fluticasone, respectively. These data confirm that K_Ca3.1 channels are essential in the regulation of TNFα/IFNγ-inducible GC-resistant chemokine in ASM cells.

K_Ca3.1 channel inhibition suppresses TNFα/IFNγ-induced steroid-resistant chemokine mRNA expression

Quantitative PCR analyses were performed to determine whether K_Ca3.1 channel blockers modulate TNFα/IFNγ-induced chemokine expression at the transcriptional level. In cells exposed to TNFα/IFNγ for 6 h, CCL5, CX3CL1 and CCL11 mRNA expression were increased significantly by 35.2 ± 19 fold, 96 ± 41 fold and 1.8 ± 0.43, respectively (P<0.05). Fluticasone did not alter on the expression of CCL5 or CCL11 mRNA, but there were non-significant trends for reduced expression of CCL5 and CCL11 mRNA with K_Ca3.1 blockade (Fig.9). Combining fluticasone with TRAM-34 or ICA-17043 led to a drastic inhibition of both CCL5 and CCL11 expression (P<0.05, n=4 combined from 2 asthmatics, 1 control, and 1 COPD)(Fig. 9A–9C).

Figure 9. K_Ca3.1 blockers modulate the transcription of GC-resistant chemokines.

Cells pre-treated with TRAM-34 (200 nM) or ICA-17043 (60 nM) with or without fluticasone (100 nM, 2 h) were stimulated with TNFα/IFNγ for an additional 4 h. Total RNA was extracted and real time quantitative PCR was performed as described in the methods section. Results are expressed as fold induction of chemokine expression by calculating the negative inverse of the 2^−ΔΔCT value for each condition. Expression levels of CCL5 (A), CX3CL1 (B) and CCL11 (C) were expressed as % of TNFα/IFNγ response. Means ± S.E.M of experiments performed in duplicate in 4 different subjects (2 asthmatics, 1 control and 1 COPD). *P < 0.05, **P<0.01 compared to TNFα/IFNγ/DMSO control.

Fluticasone alone had no significant effect on TNFα/IFNγ-induced CX3CL1 expression in response to TNFα/IFNγ, but this expression was almost completely reduced in cells treated with either K_Ca3.1 blockers alone, irrespective of fluticasone treatment (Fig. 9B). These data demonstrate that K_Ca3.1 channel inhibition regulates the expression of GC-resistant chemokines in ASM cells at the transcriptional level.

Fluticasone-induced GRα phosphorylation at ser211 and Glucocorticoid-Induced Leucine Zipper (GILZ) expression is impaired in steroid-resistant states but restored in the presence of K_Ca3.1 channel inhibitors

Cytokine-induced steroid resistance in human ASM cells involves the inhibition of GRα transcriptional activity via multiple mechanisms (12). Here, we confirmed that fluticasone-induced GRα phosphorylation on serine 211, which is essential for its transcriptional activity, was almost completely inhibited in the presence of TNFα/IFNγ (Fig.10A–B). This reduced GC-induced GR phosphorylation induced by TNFα/IFNγ was associated with an impaired expression of the well-defined GRE-dependent gene called GILZ (Fig.10C). Although these data were generated in cells form healthy subjects, similar findings were also observed in cells derived from n=2 non-severe asthmatics (data not shown). More importantly, in the presence of the K_Ca3.1 channel blockers ICA-17043 or TRAM-34, the inhibitory effect of TNFα/IFNγ on both fluticasone-induced GRα phosphorylation and GILZ expression was completely prevented (n=3, P<0.01) (Fig.10A–B and C). This demonstrates that K_Ca3.1 channel inhibition restores a key signalling component required for normal GRα transcriptional function in this ASM model of steroid resistance.

Figure 10. Impaired fluticasone-induced GRα phosphorylation and GRE-dependent Glucocorticoid-induced leucine zipper (GILZ) expression by cytokines is restored by K_Ca3.1 inhibitors.

Cells pre-treated with TRAM-34 (200 nM) or ICA-17043 (60 nM) with or without fluticasone (100 nM, 2 h) were stimulated with TNFα/IFNγ for 4 h. (A) Total cell lysates were prepared and assayed for total GR, phosphoserine 211 GR antibodies and β-actin for loading by immunoblot analysis. (B) Means ± S.E.M of scanning densitometric analyses of blots from n=3 healthy patients with each value normalized over the mean density of the corresponding total GR bands. *p<0.05 compared to fluticasone, *p<0.05 compared to TNFα/IFNγ. (C) RNA was extracted and purified using the PureLink® RNA Mini Kit according to the manufacturer's instructions. Total mRNA (2 μg) was subjected to real time RT-PCR with GILZ and β-actin primers and the relative quantification in each condition was performed using the standard curve method and expressed as fold increased over basal. Each experiment was performed in duplicate and repeated in cells from three different healthy donors. **p<0.01 compared to fluticasone alone or TNFα/IFNγ/fluticasone.

K_Ca3.1 channel inhibition suppresses cytokine-induced PP5 expression

We have recently uncovered that TNFα/IFNγ-induced upregulation of the protein phosphatase PP5 is the key factor by which this cytokine combination impairs GC-induced GRα phosphorylation at ser211 (43, 47). Flow cytometry assays confirmed that PP5 was indeed up-regulated following treatment with TNFα/IFNγ in 9 subjects (4 asthmatics and 5 healthy subjects) but interestingly, this response was significantly inhibited by both K_Ca3.1 blockers (Fig.11). This suggests that while K_Ca3.1 channel activity directly regulates the expression of CX3CL1 (Fig.6), it is also involved in the induction of protein phosphatase PP5 which in turn promotes GC resistance by dephosphorylating GRα at ser211.

Figure 11. PP5 up-regulation by TNFα/IFNγ is inhibited by K_Ca3.1 inhibitors.

Cells pre-treated with TRAM-34 (200 nM) or ICA-17043 (60 nM) for 2 h were stimulated with TNFα/IFNγ for 22 h. PP5 levels assessed by flow cytometry were expressed as the fold increase in mean fluorescence intensity (MFI) over basal ± S.E.M. of experiments performed in 9 subjects (n=4 asthmatics and n=5 healthy controls). *P <0.05; **P < 0.01.

DISCUSSION

K_Ca3.1 channels may represent important players in the pathogenesis of several lung diseases (19). The present article now reveals that K_Ca3.1 channels may play an important role in the regulation of GC insensitive pathways. We found that TNFα/IFNγ-induced expression of GC-resistant CX3CL1 was inhibited by K_Ca3.1 channel blockade or knockdown at both the mRNA and protein levels, irrespective of GC presence. In contrast, the inhibition of GC-resistant CCL5 was only visible in cells co-treated with both fluticasone and K_Ca3.1 channel inhibition. The failure of fluticasone to induce GRα phosphorylation at ser211 or to promote the expression of the GC-inducible gene GILZ in the GC-resistant state was fully prevented by the presence of K_Ca3.1 inhibitors as a consequence of reduced PP5 expression. This is the first report in any given cell type or tissue that describes a functional interaction between K_Ca3.1 channels and GC signalling.

The mechanisms underlying GC resistance in asthma are not clearly defined (48). However, combined in vitro and in vivo studies have led to the important conclusion that GC-resistant pathways are operative in ASM cells within asthmatic airways (49). Indeed, many studies performed on bronchial biopsies from asthmatics showed marked expression of inflammatory proteins such as CX3CL1 (13), CCL11 (14), CCL15 (15), CCL19 (16), ADAM33 and ADAM8 (17, 18) within ASM bundles despite patients being treated with either high dose inhaled or oral GC. We have previously modeled this GC-resistant state in vitro by exposing cultured human tracheal ASM cells to a combination of TNFα/IFNγ which results in the production of an array of inflammatory proteins including CD38, CCL5, CX3CL1 and IRF-1 that are completely resistant to GC treatment (12, 43, 47, 50, 51). Here, we found that in addition to CCL5 and CX3CL1, production of other chemokines, CCL11 and CXCL10 are also insensitive to fluticasone treatment; more importantly, this GC-resistant state also occurs in primary bronchial ASM cells in health, asthma and COPD (as opposed to tracheal ASM cells in our previous studies). Generally, TNFα/IFNγ-dependent steroid-resistant chemokine production was similar irrespective of disease status (healthy, COPD, asthma), although production of CX3CL1 was dramatically reduced by more than 80% in COPD patients when compared to levels produced by cells from healthy subjects. The reasons for this are unknown and additional studies are clearly required to confirm this observation.

Immunohistochemistry on bronchial biopsies demonstrated that K_Ca3.1 channels are expressed in vivo in ASM bundles in asthmatic patients, and interestingly, this included a nuclear distribution. This observation was further confirmed by assessing K_Ca3.1 expression in cultured ASM cells using immunoblot analysis directly on nuclear extracts and immunofluorescence staining showing a nuclear distribution of K_Ca3.1. K_Ca3.1 expression in asthmatic ASM bundles was not affected by disease severity and the associated intensity of anti-asthma treatment, and no differences in immunostaining intensity were evident between healthy subjects and asthmatic patients. This suggests that changes in channel activity rather than protein expression within ASM bundles could explain their contribution to the pathogenesis of asthma. In contrast to the study showing that TGFβ increased both K_Ca3.1 expression and activity in ASM cells (20), we found that TNFα/IFNγ failed to significantly stimulate K_Ca3.1 protein levels, or plasma membrane K_Ca3.1 channel activity, suggesting that modulation of K_Ca3.1 expression/function in ASM is highly stimulus-dependent (growth factors versus cytokines). The fact that there is no change in K_Ca3.1 activity when measured at the cell surface supports the concept that compartmental changes in Ca²⁺ levels rather than cytoplasmic changes could explain the activation of K_Ca3.1 in GC sensitivity. Our data showing that K_Ca3.1 inhibition does not affect CXCL10 induction strongly suggest that the inhibitory effect of K_Ca3.1 blockade is gene-specific and not due to a generalized non-specific effects linked to overall changes in Ca²⁺ levels inside the cells. Here we also confirmed our previous reports (20) showing that K_Ca3.1 inhibitors had no cytotoxic action on ASM cells. Because expression of K_Ca3.1 channels were found to be expressed in the nuclear compartment, our study raises the possibility that nuclear modulation of K_Ca3.1 function by TNFα/IFNγ could explain their involvement in modulating GC signalling. Interestingly, the expression of K_Ca3.1 channels has been reported in cytoplasmic organelles such as mitochondria in a human colon tumour cell line (45). Additional studies are clearly needed to define how intracellular K_Ca3.1 channels regulate cellular function in ASM cells.

We found that combining fluticasone with K_Ca3.1 blockers or K_Ca3.1 downregulation was effective in inhibiting GC-resistant CCL5, CCL11 and CX3CL1 in ASM cells. It is interesting to note that while K_Ca3.1 blockade had a strong inhibitory effect on cytokine-induced CCL5 expression at the mRNA level, production of CCL5 protein was only reduced by 50%. We believe that in addition to transcriptional mechanisms, CCL5 induction by cytokines is also regulated at the post-transcriptional level. Previous reports performed in A549 cells (52) and in human ASM cells (53) support this hypothesis by showing that while IFNγ on its own did not stimulate CCL5 expression, it does enhance TNFα-induced CCL5 expression via post-transcriptional mechanisms. We also have some preliminary evidence using a different pharmacological inhibitor demonstrating that the degree of CCL5 inhibition seen at the mRNA levels does not correlate with similar changes at the protein level (Amrani, unpublished observation).

The ability to reproduce the same results when channels were downregulated by shRNA shows unequivocally that K_Ca3.1 regulates, at least in part, TNFα/IFNγ-dependent GC-resistant chemokine expression. In our study, we noted some expected discrepancy between the two inhibitory strategies. For example, cytokine-induced CCL5 expression was significantly affected by K_Ca3.1 shRNA but not by the pharmacological blockers and the degree of CX3CL1 inhibition was somewhat greater when combining fluticasone with pharmacological blockers and not with shRNA vectors. In contrast to soluble inhibitors, channel knockdown could indirectly impact on other cellular signalling pathways through the loss of interactions between the target protein (here the K_Ca3.1 channel) and other key binding partners. In the case of K_Ca3.1, very little is known about the nature of proteins that associate with the channel. The direct binding of 5'-AMP-activated protein kinase (AMPK) (54), mammalian protein histidine phosphatase (PHPT-1) (55), or nucleoside diphosphate kinase B (NDPK-B) to K_Ca3.1 was found to be essential in regulating channel activity (56). Considering the fact that these proteins have multifunctional properties, it is therefore plausible that reducing K_Ca3.1 levels could lead to downstream effects not evoked by channel blockers alone. The apparent differences observed between the data obtained with the soluble inhibitors and silencing adenoviruses could also be due to the heterogeneity in patients' responses since different subjects were used with the two inhibitory strategies.

Interestingly, GC-resistant CX3CL1 expression, at both protein and mRNA levels, was inhibited by K_Ca3.1 inhibitors irrespective of GC treatment, suggesting that K_Ca3.1 channel activity is a key factor regulating the transcriptional expression of this chemokine in response to TNFα/IFNγ. The putative role of K_Ca3.1 in regulating expression of inflammatory mediators has been mostly described in T lymphocytes where K_Ca3.1 blockade (either via pharmacological inhibitors or the use of T cells deficient in the channel) significantly reduces T cell receptor-induced expression of IL-2, TNFα and IFNγ (57–59). In T cells it is likely that this results from the instrumental role of K_Ca3.1 in regulating Ca²⁺ entry through the plasma membrane, a vital signal for optimal T cell activation and cytokine secretion. The fact that TNFα/IFNγ-dependent CX3CL1 expression in ASM cells is inhibited by two selective blockers of the K_Ca3.1 pore demonstrates that the ionic conductance of K⁺ is key in the mediation of this effect, and not a regulatory channel domain. Interestingly, we and others have shown that TNFα alters Ca²⁺ handling in ASM cells via the induction of ectoenzyme CD38 (60, 61) or transient receptor potential C3 channels (62). It is therefore plausible that K_Ca3.1 regulates CX3CL1 via activation of Ca²⁺ Ca²⁺-dependent pathways. However, as discussed above, it seems unlikely that this is due to channel activity located in the plasma membrane.

Another major observation in our study is the functional interaction between K_Ca3.1 channels and GC signalling pathways. Specifically, K_Ca3.1 inhibitors restored the ability of fluticasone to inhibit the production of CCL5 and CCL11 in TNFα/IFNγ-induced GC-resistant conditions, irrespective of the K_Ca3.1 inhibition strategy used. In addition, this restoration of cell sensitivity to fluticasone by K_Ca3.1 blockade was concomitantly associated with a reinstatement of the GRα phosphorylation at ser211 that was impaired in TNFα/IFNγ-treated cells. This indicates that K_Ca3.1 activity drives TNFα/IFNγ-induced GC insensitivity in ASM cells, in part by via the modulation of GRα phosphorylation status (Fig.10A) and GRα transactivation activity (Fig.10B). Our data support the growing evidence showing the importance of transactivation properties in the anti-inflammatory action of GC (63, 64). Although GRα phosphorylation on three major residues located on its N terminus (ser203, ser211, and ser226) affects key functions of the receptor including turnover and subcellular trafficking, it is phosphorylation on ser211 that is essential for optimal GRα transcriptional activity (65). Indeed, in agreement with our previous reports (12, 47, 50, 51), we confirmed that fluticasone-induced GRα-dependent transactivation was significantly impaired in the TNFα/IFNγ-induced GC-insensitive state (Fig.10C). More importantly, K_Ca3.1 blockade was also able to fully restore fluticasone-induced GRα-dependent transactivation, suggesting a role of K_Ca3.1 channel in impairing GC function in ASM cells. Our previous report showed that the impaired GRα transactivation was due to the up-regulation of the serine/threonine phosphatase PP5 which mediated cytokine-induced GRα dephosphorylation at ser211 (43). Here, we not only confirm that PP5 levels are increased by TNFα/IFNγ in ASM cells independently of disease status (healthy and asthmatics), but more importantly, that PP5 induction was dependent on functional K_Ca3.1 channels. Although a direct link between PP5 and steroid insensitivity has been suggested by Goleva and colleagues who found that PP5 knockdown restored GC responsiveness in oestrogen-treated breast cancer cells (66), our report is the first to show that this functional link between pro-asthmatic cytokines and PP5 occurs via K_Ca3.1-dependent pathways. Our present study supports the novel model that TNFα/IFNγ impairs GC sensitivity in ASM cells by promoting GRα dephosphorylation via K_Ca3.1-dependent up-regulation of PP5 expression.

In summary, we have shown that K_Ca3.1 channels contribute to TNFα/IFNγ-associated GC-insensitivity (Fig.12). Transcription of some pro-asthmatic genes such as CX3CL1 is driven by K_Ca3.1-dependent pathways that are insensitive to corticosteroids. Other genes such as CCL5 or CCL11 are indirectly rendered resistant to GC via the induction of the K_Ca3.1-dependent serine/threonine phosphatase PP5 which interferes with GRα receptor function through decreased GRα phosphorylation and transactivation. Our study uncovers the potential therapeutic value of targeting the K_Ca3.1-PP5 axis in the treatment of lung diseases such as severe asthma where relative GC resistance is evident. The availability of a well-tolerated orally bioavailable K_Ca3.1 blocker that has been used in phase III trials of sickle cell disease (ICA-17043 [Senicapoc]) (19) means that there is the potential for the rapid translation of these findings to the clinic.

Figure 12. Role of K_Ca3.1 in mediating GC insensitivity in ASM cells.

We uncovered two mechanisms by which K_Ca3.1 channels mediate TNFα/IFNγ-associated corticosteroid insensitivity. Transcription of some pro-asthmatic genes involves K_Ca3.1-dependent pathways that are insensitive to corticosteroids (right). Other genes are resistant to corticosteroids via the induction of the K_Ca3.1-dependent serine/threonine phosphatase PP5 which impairs GRα transcriptional function through decreased GRα phosphorylation (left).

Abbreviations

ASM

Airway smooth muscle

ADAM

A Disintegrin And Metalloprotease domain

GC

Glucocorticoid

CCL

Chemokine (C-C motif) ligand

CX3CL

Chemokine (C-X3-C motif) ligand

COPD

Chronic Obstructive Pulmonary Disease

GR

Glucocorticoid receptor

GRIP-1

Glucocorticoid receptor interacting protein 1

GILZ

Glucocorticoid-Induced Leucine Zipper

IRF-1 Protein

Interferon regulatory factor-1

PP5

Phosphatase 5

shRNA

Small hairpin RNA

TLSP

Thymic stromal lymphopoietin