Roflumilast Partially Reverses Smoke-induced Mucociliary Dysfunction

Andreas Schmid; Nathalie Baumlin; Pedro Ivonnet; John S. Dennis; Michael Campos; Stefanie Krick; Matthias Salathe


Respiratory Research. 2015;16(135) 

In This Article


Effect of Roflumilast on Intracellular cAMP

To evaluate the effect of roflumilast on intracellular cAMP concentration ([cAMP]i), we used our previously described FRET assay of fluorescently tagged PKA subunits[13] where changes in FRET ratio (ΔFRET ratio) are indicative of changes in [cAMP]i.[13] Apical addition of 10 μM forskolin as well as 20 μM albuterol increased [cAMP]i in NHBE cells. While roflumilast did not increase cAMP when added alone, it increased the peak [cAMP]i response to forskolin and albuterol (Fig. 1a) and prolonged the [cAMP]i increase after forskolin washout (Fig. 1b).

Figure 1.

Effect of roflumilast on real-time [cAMP]i, estimated by FRET with and without smoke. a Treatment with 100 nM roflumilast alone (10 min) does not increase [cAMP]i, but roflumilast enhances forskolin- and albuterol-mediated [cAMP]i increases (10 μM forskolin and 20 μM albuterol in control cultures without airflow exposure (white bars; forskolin three lungs, n = 21; albuterol one lung, n = 8). b Representative tracings of FRET ratios over time as a reflection of changes in [cAMP]i upon stimulation with forskolin in the presence (red) and absence (black) of roflumilast. c Smoke exposed cells (black bar) show a decreased baseline FRET ratio compared to cultures exposed to airflow (grey bar), possibly indicating a decreased intracellular cAMP level (three different lungs, n = 30). However, FRET ratios under these conditions cannot be calibrated to [cAMP]i. d Roflumilast leads to a significant increase in the forskolin-induced ΔFRET ratio in air control (gray) and smoke exposed (black) cultures using the Vitrocell VC-10 smoke exposure system (three different lungs, n = 21–43). e Roflumilast treated cultures show a prolonged elevation of the FRET ratio after washout of forskolin in air control and smoke exposed cells (three different lungs, n = 21–43). *p < 0.05

Roflumilast also increased the [cAMP]i peak response and duration to 10 μM forskolin in cultures exposed to smoke or air (air control) in the Vitrocell VC-10 smoking robot. In smoke-exposed cells, the agonist-induced peak ΔFRET ratio response was significantly less compared to the air exposed control cells in the presence of roflumilast (Fig. 1d; each data point from three lungs with n ≥ 21 measurements), while no difference was found for the prolonged cAMP increase between smoke and air control (Fig. 1e). To assess the effect of cigarette smoke on intracellular cAMP concentrations, we compared baseline FRET ratios of smoked and air control cultures (Fig. 1c). We found that cultures exposed to cigarette smoke had a significantly lower baseline FRET ratio compared to air control cultures (0.958 ± 0.023 vs 0.980 ± 0.015; 3 lungs, n = 30, p < 0.05). While calibrations of these baselines are not possible and these results have to take this caveat into account, the data could indicate that acute cigarette smoke exposure decreased intracellular cAMP concentrations. On the other hand, it did not decrease changes in [cAMP]i responses to agonist stimulation and brought FRET ratios in smoked cells to the baseline FRET ratios of control cells in the absence of roflumilast. The further augmenting effect of roflumilast on peak forskolin responses, however, was reduced by cigarette smoke. On the other hand, roflumilast still significantly prolonged the increase in [cAMP]i in smoked cells after forskolin washout.

Effect of Roflumilast on Apical Chloride Conductance

The effect of roflumilast on apical chloride conductance was measured by short circuit currents (Isc) in NHBE cells. Adding 100 nM roflumilast alone (Fig. 2a) resulted in a minimal current change of 0.99 ± 0.22 μA/cm2 (14 lungs, n = 29), which was significantly less than the changes observed after apical addition of 10 μM albuterol (11.75 ± 2.62 μA/cm2, p < 0.05). Use of a chloride-free apical solution did not enhance the effect of roflumilast (not shown).

Figure 2.

Effect of roflumilast on apical chloride conductance with and without smoke exposure. a Evaluation of chloride conductance in control cultures without airflow exposure (white bars): Application of roflumilast (100 nM) alone induces a minimal increase in Isc (14 lungs, n = 29). Albuterol (10 μM) induces an increase in peak chloride efflux (ΔIsc) that is augmented by pre-treatment with roflumilast (100 nM) (six to ten lungs; n = 6–10). b Representative tracing of an Ussing chamber experiment demonstrates an increase of chloride efflux after administration of 10 μM of albuterol (black). Application of roflumilast alone hardly increases the conductance, whereas the addition of albuterol to roflumilast (red) treated cultures significantly increases chloride conductance compared to untreated cultures. c Chloride efflux is significantly decreased in cultures exposed to cigarette smoke (black bars) using the Vitrocell VC-10 smoking robot when compared to air control (grey bars). Roflumilast significantly increases apical chloride efflux in air-exposed cells and rescues smoke-associated decreases in chloride conductance (19 lungs; n = 19). d Control experiments, DMSO (vehicle) does not show any changes in the response of chloride conductance to albuterol in air and cigarette smoke-exposed cultures (two lungs; n = 4). *p < 0.05

Pretreatment with roflumilast significantly increased the response to albuterol (19.02 ± 2.27 μA/cm2; p < 0.05). The dynamics of chloride conductance in cultures treated with roflumilast, albuterol and their combination is demonstrated in Fig. 2b. Effects of 100 nM roflumilast on cultures exposed to control airflow or cigarette smoke in the Vitrocell VC-10 smoking robot are shown in Fig. 2c. Roflumilast increased the short circuit current response to albuterol in air-exposed control cells (6.69 ± 1.14 μA/cm2 vs 4.8 ± 0.71 μA/cm2; p < 0.05). Smoke exposure significantly decreased chloride conductance compared to controls (2.1 ± 0.26 μA/cm2 vs 4.8 ± 0.71 μA/cm2; p < 0.05) but roflumilast reversed the smoke effect (5.26 ± 0.65 μA/cm2 vs 2.1 ± 0.26 μA/cm2; p < 0.05) so that changes in Isc in roflumilast-treated, smoked-exposed cultures were similar to untreated airflow controls (5.26 ± 0.65 μA/cm2 vs 4.8 ± 0.71 μA/cm2; p > 0.05). As shown in Fig. 2d, DMSO did have no effect on chloride conductance in air control (9.4 ± 1.5 μA/cm2 vs 8.9 ± 2.3 μA/cm2) and cigarette smoke exposed cells (4.1 ± 0.6 μA/cm2 vs 3.9 ± 0.4 μA/cm2 : 2 different lungs, n = 4).

These data show that roflumilast increased albuterol-stimulated apical chloride efflux and rescued the negative effect of cigarette smoke on this conductance. Addition of roflumilast without additional stimulation with albuterol or forskolin increased Isc only minimally.

Effect of Roflumilast on CFTR Function in Smoke-exposed Cells

To assure that the observed effects of roflumilast on Isc were truly related to Cl efflux via CFTR, smoke- or air-exposed ALI cultures were pretreated basolaterally with 100 nM roflumilast, mounted in Ussing chambers, stimulated with 10 μM albuterol and then 10 μM CFTR antagonist CFTRinh 172 was added apically (Fig. 3a). The observed decrease in Isc is a more specific measure of CFTR activity. Roflumilast increased CFTR-dependent Cl secretion in air-exposed cells (−6.74 ± 1.79 μA/cm2 vs −4.81 ± 1.29 μA/cm2; p < 0.05, n = 19). Smoke exposure decreased CFTR conductance compared to air control (−2.26 ± 0.51 μA/cm2 vs −4.51 ± 1.29 μA/cm2; p < 0.05) and roflumilast improved CFTR function in smoked cells to a level similar to controls (−3.54 ± 1.00 μA/cm2 vs −4.51 ± 1.29 μA/cm2; p < 0.05; Fig. 3a). The ratio of CFTR conductance in response to albuterol in the absence and presence of roflumilast was not statistically different between smoke- and air-exposed control cells (1.34 ± 0.19 vs 1.9 ± 0,37; p > 0.05; Fig. 3b). These data show that roflumilast enhanced impaired CFTR function in NHBE cell cultures exposed to cigarette smoke.

Figure 3.

Effect of roflumilast on CFTR function. a Basolateral pretreatment of cells exposed to either air flow (air control; grey bars) or cigarette smoke (black bars) with 100 nM roflumilast improves CFTR function as measured by the decrease in Isc after apical application of 10 μM CFTRinh 172; all cells were stimulated with 10 μM albuterol prior to CFTR inhibition. Roflumilast rescues CFTR function in smoke exposed cells to a level not different to the untreated air exposed cells. b Air and cigarette smoke exposed cultures show similar fold increases in CFTR conductance with and without roflumilast (19 lungs; n = 19). * p < 0.05

Effect of Roflumilast on Airway Surface Liquid (ASL)

To measure airway surface liquid (ASL) volume, a previously described meniscus scanning method[20] was used as outlined in methods. A good correlation of measured volumes with the meniscus scanning method was found when measuring 30 min and 3 h after addition of 5, 10 and 20 μl PBS to control cells (Fig. 4a). ASL measurements were taken at different time points after smoke or air exposure (Fig. 4b and c). Baseline measurements were taken 1 h after exposure, just prior to roflumilast addition (100 nM; all n = 6). Measurements of ASL were done after 4, 7, and 24 h of smoke/air exposure. While air-exposed cells increased ASL over time after exposure to reach baseline again after 24 h, smoke-exposed cells revealed a blunted response with a decrease of ASL that trended to fall below baseline after 24 h (however this did not reach statistical significance). Roflumilast enhanced ASL volume increases in both air- and smoke-exposed cells (Fig. 4b and c). At 24 h, the roflumilast treated, smoke exposed cells had an ASL volume that was significantly higher than the control cells that were only smoke exposed.

Figure 4.

Effect of roflumilast on Airway Surface Liquid (ASL). a Calibration of ASL measurements shows a good correlation between added and measured Δ volumes in 12 mm Corning Transwell filters measured 30 min after addition (n = 7). Similar results were found when measuring after 3 h (not shown). b Following ASL volumes over time after exposure to air shows an initial increase in ASL volume that returns to baseline after 24 h. Roflumilast enhances the response and keeps ASL above baseline at the 24 h time point (six different lungs, n = 6, # p < 0.05 for comparison between control and roflumilast treated cells at different time points). c Following ASL volumes over time after exposure to cigarette smoke shows a blunted initial increase. Roflumilast on the other hand, increases initial volume and restores the ASL volume response after smoke exposure, comparable to air exposure, keeping the ASL volume close to baseline at 24 h (six different lungs, n = 6, # p < 0.05 for comparison between control and roflumilast treated cells at different time points). Black and red dotted line in B/C for comparison of baseline of control (black) and roflumilast treated (red) cultures. d Control experiments with DMSO as vehicle do not show any difference in Δ ASL measurements between 1 and 4 h in air versus cigarette smoke exposed cultures (two lungs; n = 4). * p < 0.05

These data show that roflumilast enhances ASL volume in NHBE cell cultures and partially reverses the negative effect of cigarette smoke.

Ciliary Beat Frequency (CBF)

To examine the effect of roflumilast on smoke-mediated changes in CBF, fully differentiated NHBE cells were used after basolateral treatment with 100 nM roflumilast. DMSO controls were published by us before: DMSO did not show any significant effects on CBF.[13] Baseline measurements were made one hour after roflumilast addition, before the cells were exposed to cigarette smoke or air in the Vitrocell VC-10 robot (control). Three hours after smoke/air exposure (four hours after treatment with roflumilast), CBF was measured. Baseline CBF of the cultures was 4.5 ± 0.2 Hz before and 4.68 ± 0.2 Hz 1 h after exposure to 100 nM roflumilast (p > 0.05; five lungs, n > 14). Exposure of the cultures to airflow (air control) did not change CBF in untreated cells. However, airflow increased CBF significantly when pre-treated with roflumilast for 4 h (4.19 ± 0.25 Hz vs. 6.63 ± 0.5 Hz; p < 0.05). Three hours after exposure to cigarette smoke, CBF decreased compared to air control (1.28 ± 0.06 Hz vs 4.19 ± 0.24 Hz; n > 14 each; p < 0.05) and non-exposed cultures (4.5 ± 0.2 Hz). Roflumilast increased CBF of the smoke-exposed cells significantly at 3 h (3.31 ± 0.43 Hz vs 1.28 ± 0.06 Hz; p < 0.05) and CBF of roflumilast treated, smoke-exposed cultures was not different compared to air-control cultures (3.31 ± 0.43 Hz vs 4.19 ± 0.24 Hz, p > 0.05). After initial CBF measurements, 50 μl PBS was added apically to replenish ASL volume. This maneuver equalized CBF in air- and smoke-exposed cells with or without roflumilast (5.78 ± 0.54 Hz vs 6.67 ± 0.22 Hz vs 6.04 ± 0.3 Hz vs 6.82 ± 0.37 Hz; p > 0.05 for all comparisons; Fig. 5a), possibly indicating ASL volume loss as the main reason for the CBF decrease rather than a direct smoke effect on cilia. However, a lower [cAMP] after smoke exposure may contribute here as well.

Figure 5.

Effect of roflumilast on ciliary beat frequency (CBF). a CBF measurement after roflumilast exposure for 4 h (baseline in white bars on the left of the graph after 1 h; five lungs; n = 14–40). Roflumilast-treated cultures did not show an increase in CBF compared to untreated cultures (baseline). Smoke (4 h after roflumilast) significantly decreases CBF 3 h after exposure, an effect reversed by roflumilast. Exposure to air (air control) increases CBF upon administration of roflumilast (gray bars). Bars on a blue background indicate measurements of CBF after rehydration of the apical surface with 50 μl PBS. CBF of cells exposed to smoke vs air equalizes, indicating a role of ASL volume depletion in the CBF changes (* p < 0.05). b CBF measurements after short-term exposure of roflumilast for 15 min followed by addition of 10 μM forskolin in a submerged, two chamber perfusion system (three lungs; n = 10-14). In air control cultures, forskolin increases CBF, but the addition of roflumilast does not further enhance CBF. In smoke-exposed cells, CBF increases upon forskolin addition, and again with roflumilast. CBF baselines are low in smoke-exposed cells, possibly correlating to the lower baseline cAMP levels in smoke exposed cells as assessed by FRET (taken at the same time points). * p < 0.05

To evaluate the effect of roflumilast on CBF independent of ASL volume, we measured CBF in submerged conditions. Treatment with 10 μM forskolin increased CBF in air control cells in the absence and presence of roflumilast (8.75 ± 0.5 Hz vs 10.15 ± 0.55 Hz and 8.73 ± 0.6 Hz vs 10.9 ± 0.62 Hz respectively; n > 10; p < 0.05; Fig. 5b). Similar changes were seen in smoke-exposed cells and smoke-exposed cells treated with roflumilast (4.48 ± 0.31 Hz vs 5.69 ± 0.33 Hz and 5.29 ± 0.25 Hz vs 7.29 ± 0.29 Hz respectively; p < 0.05). CBF of smoke-exposed cells treated with roflumilast and forskolin had a higher CBF than cells treated only with forskolin (7.29 ± 0.29 Hz vs 5.69 ± 0.33 Hz; p < 0.05). This indicates that roflumilast had a stimulatory effect on smoke-exposed cilia directly, independent of ASL volume changes. However, the same phenomenon was not observed in air control cells.

The increased baseline of CBF in these submerged experiments compared to the other CBF measurements after apical PBS addition may be related to the fact that the evaluations were done with constant fluid flow in the apical compartment that may stimulate CBF. These data suggest that the CBF changes in Fig. 5b are not related to ASL volume changes, but rather to an increase in cAMP production (see Fig. 1) directly affecting CBF.

In summary, roflumilast has a dual effect on CBF: it directly reverses smoke-induced CBF decreases and increases smoke-related reductions in ASL volume, thereby indirectly allowing normal ciliary beating. The smoke-induced lower CBF baselines in Fig. 5a and b may be due to the lower baseline cAMP levels after smoke exposure suggested by the FRET experiments discussed above.

Effects of the Long Acting ß2 Adrenergic Agonist Formoterol With Roflumilast

All experiments so far described results using forskolin or the short acting ß2 mimetic albuterol to increase cAMP levels. In clinical practice, long acting ß2 adrenergic agonists are used for chronic treatment of COPD, whereas short acting ß2 agonists are employed as rescue inhalers for symptomatic relief.[21] Based on this fact, we examined the effect of formoterol on CFTR and CBF in the absence and presence of roflumilast in control and smoke exposed cultures.

Figure 6a shows specific CFTR conductance evaluated by Isc decreases upon channel inhibition with CFTRinh 172 in the acute presence of 10 μM albuterol in addition to the experimental conditions. Cultures were incubated with 100 mM roflumilast, 100 nM formoterol, both or none for 2 h before adding 10 μM of albuterol. Specific CFTR conductance was significantly decreased in cigarette smoke exposed compared to control cultures (−3 ± 0.79 vs −6.9 ± 1.99 μA/cm2). Roflumilast alone did not significantly improve CFTR function either in smoked (−4.58 ± 1.67 μA/cm2) or in control cultures (−9.6 ± 2.8 μA). Formoterol alone, however, improved CFTR function (−6.37 ± 1.2 μA/cm2 and −12.6 ± 2.3 μA/cm2 respectively; 11 lungs, n = 11). Next, we evaluated the effect of the combination of formoterol and roflumilast on CBF (Fig. 6b). As expected, CBF was decreased in cigarette smoke exposed versus control cells (2.5 ± 0.7 Hz vs 8.2 ± 0.6 Hz; n = 11; p < 0.05). Neither roflumilast nor formoterol alone improved CBF. However, the combination of the both significantly increased CBF in smoked cells to a level similar to control cultures (7.9 ± 0.6 vs 7.9 ± 0.6 Hz; 11 lungs, n = 11).

Figure 6.

Effect of long acting ß2 mimetic in combination with roflumilast on CFTR function and ciliary beat frequency (CBF). a CFTR function, assessed by adding 10 μM albuterol followed by CFTRinh 172, shows a significant decrease in cigarette smoke-exposed cell cultures compared to air control conditions. CFTR function remains significantly lower in smoke exposed cells compared to air controls when adding roflumilast, formoterol or both. Treatment with formoterol improves CFTR function in both smoke and air control compared to non-treated cells, whereas treatment with roflumilast alone under the same conditions does not. b CBF is decreased in smoke exposed cells compared to air controls. Neither roflumilast nor formoterol alone rescue smoke-induced decreases in CBF. However, the combination of both restores CBF levels to air control. * for p < 0.05, 11 lungs with n = 11 for all experiments

CFTR Expression in Cells Brushed From Patients

We examined the effect of cigarette smoke and roflumilast on CFTR mRNA expression in airway epithelia cells (Fig. 7). First, qPCR was performed on cells from bronchial brushes obtained during bronchoscopies of individuals with different smoking histories (Fig. 7a). The groups were patients that never smoked, current smokers (at least 10 pack years), and ex-smokers (quit smoking at least 1 year before collection of the cells). We found that current smokers had a significant increase in CFTR expression compared to non-smokers and ex-smokers (44.5 ± 8.1 vs 16.4 ± 2.3 vs 15.3 ± 2.1; expression relative to GAPDH * 1000; each n > 12; p < 0.05). CFTR expression in ex-smokers returned to the level of non-smokers (p > 0.05).

Figure 7.

Change in CFTR expression in brushed cells obtained during bronchoscopies and NHBE cultures. a Expression of CFTR is increased in brushed cells obtained from airways of smokers compared to non-smokers and ex-smokers. (four to six lungs, n = 12–18). b In cultured epithelial cells from non-smokers, acute exposure to cigarette smoke does not change CFTR expression compared to control air-exposed cells within 3 h of exposure. Roflumilast (red framed boxes) appears to increase CFTR expression in the smoke-exposed cells after 3 h (four lungs, n = 12) * for p < 0.05

CFTR expression in cultured airway epithelial cells from non-smokers (Fig. 7b) did not increase acutely after exposure to cigarette smoke (four cigarettes) compared to air control (0.89 ± 0.2 vs 0.88 ± 0.1; p > 0.05). When air control cultures were treated with 100 nM roflumilast, there was also no significant change in CFTR expression (0.88 ± 0.1 vs 1.08 ± 0.1; n = 12; p > 0.05). However, when cells from non-smokers, exposed to smoke were treated with roflumilast, an increase in CFTR expression was observed (0.88 ± 0.18 vs 1.45 ± 0.2; p < 0.05).