Resveratrol restores intracellular transport in cystic fibrosis epithelial cells

Abstract

We have demonstrated previously that intracellular transport is impaired in cystic fibrosis (CF) epithelial cells. This impairment is related to both growth and inflammatory regulation in CF cell and animal models. Understanding how transport in CF cells is regulated and identifying means to manipulate that regulation are key to identifying new therapies that can address key CF phenotypes. It was hypothesized that resveratrol could replicate these benefits since it interfaces with multiple pathways identified to affect microtubule regulation in CF. It was found that resveratrol treatment significantly restored intracellular transport as determined by monitoring both cholesterol distribution and the distribution of rab7-positive organelles in CF cells. This restoration of intracellular transport is due to correction of both microtubule formation rates and microtubule acetylation in cultured CF cell models and primary nasal epithelial cells. Mechanistically, the effect of resveratrol on microtubule regulation and intracellular transport was dependent on peroxisome proliferator-activated receptor-γ signaling and its ability to act as a pan-histone deacetylase (HDAC) inhibitor. Resveratrol represents a candidate compound with known anti-inflammatory properties that can restore both microtubule formation and acetylation in CF epithelial cells.

INTRODUCTION

Previous work has identified perinuclear cholesterol accumulation due to impaired endosomal transport as a distinct phenotype in cystic fibrosis (CF) epithelial cells, a process that can be manipulated through a variety of pathways (6, 25, 27, 38, 53). Alterations to microtubule regulation, including reduced acetylation and slowed rates of formation, are directly responsible for the intracellular transport changes observed in CF cells (37). These changes to microtubule regulation and intracellular transport result in signaling cascades associated with inflammatory signaling and growth regulation in CF (20, 22, 36), so identifying easily accessible compounds that can reverse these alterations can have important consequences for therapeutic development.

Mechanisms controlling CF microtubule regulation have been identified. An early cellular event leading to reduced microtubule growth rates is the lower activity levels of the exchange protein activated by cAMP-1 (EPAC1) in CF cells (38). EPAC1 has multiple functions, but one is acting as a microtubule elongation factor (45). We have demonstrated that activation of EPAC1 leads to improved microtubule formation rates in primary CF epithelial cells, resulting in improved intracellular cholesterol distribution due to improved endosomal transport (38). Recent findings also demonstrate that ibuprofen treatment reverses the cholesterol accumulation phenotype by correction of CF microtubule dynamics through an adenosine monophosphate-activated protein kinase (AMPK)-dependent pathway (39). The efficacy of ibuprofen in regulating this pathway is significant, since multiple studies have shown that high-dose ibuprofen therapy is effective at controlling inflammation and reducing the rate of pulmonary function decline, particularly in younger CF patients (16–18, 21). However, the use of ibuprofen in the CF population is limited due to potential adverse effects, so new interventions are needed (2, 19, 24, 44). Finally, inhibition of histone deacetylase 6 (HDAC6), a deacetylase that regulates tubulin acetylation, restores CF microtubule acetylation, reestablishes intracellular transport, and reduces inflammatory signaling in CF models (37). A CF mouse model with Hdac6 expression depleted has restored growth and fertility, further demonstrating the potential therapeutic implications of targeting microtubule defects in CF (36). The goal of this study is to define other mechanisms of microtubule regulation that can be manipulated with a common compound for potential therapeutic use.

It is hypothesized in this study that resveratrol represents an appealing intervention for CF microtubule regulation based on the activation of sirtuins 1 (SIRT1) and 3 (SIRT3). SIRT1 and SIRT3 are deacetylases that regulate a number of cellular functions related to metabolism and inflammation (28, 48). Changes in CF metabolomics and mitochondrial function also point potentially for a need to activate sirtuins (12, 14, 49). Directly relevant to this study, activation of both SIRT1 and SIRT3 has been shown to promote microtubule polymerization and stability (8, 15, 42). Resveratrol also stimulates other pathways that may be relevant to microtubule stability in CF cells such as the stimulation of adenosine monophosphate-activated protein kinase (AMPK) activity (3). As previously noted, ibuprofen corrects CF microtubule dynamics through AMPK activation (39).

In this paper, the ability of resveratrol to improve intracellular transport as marked by improved endosomal trafficking through the regulation of microtubule dynamics and stability is shown. The mechanisms of action are explored, and novel regulatory interactions are defined. The use of resveratrol represents an easily accessible means of modulating microtubule regulation in CF cells and correcting aspects of CF cell biology.

METHODS

Cells.

9/HTEo⁻ cells originally developed by Dr. Dieter Gruenert were human tracheal epithelial cells that transformed by simian virus 40. They were grown at 37°C in a 95% O₂-5% CO₂ incubator on 10-cm-diameter tissue culture dishes (Corning, Corning, NY) in 10 ml Dulbecco's modified Eagle's medium (DMEM; Gibco, Carlsbad, CA) supplemented with 10% FBS containing 2 mM l-glutamine, 0.04 mM HEPES, 1 U/ml penicillin, and 1 µg/ml streptomycin. IB3 cells (CF), human bronchial epithelial cells isolated and immortalized from a CF patient with ΔF508 mutation, and S9 cells [wild type (WT)], IB3 cell mutation corrected with full-length WT cystic fibrosis transmembrane conductance regulator (CFTR), were a gift from Pamela L. Zeitlin (Johns Hopkins University, Baltimore, MD). They were grown at 37°C in a 95% O₂-5% CO₂ incubator on 10-cm-diameter tissue culture dishes in LHC-8 basal medium (Gibco) supplemented with 5% FBS containing 1 U/ml penicillin and 1 µg/ml streptomycin. Primary tissue CF mouse nasal epithelium (MNE) was excised from WT and Cftr^−/− mice. They were grown at 37°C in a 95% O₂-5% CO₂ incubator on T-75 flasks (Corning) in DMEM-Ham’s F-12 medium containing Y-27632, penicillin, and streptomycin. Human nasal epithelial cells were obtained from the CWRU CF Center cell culture core facility.

Western immunoblotting.

Acetylated-α-tubulin (mouse) antibodies were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). α-Tubulin (rabbit) antibodies were obtained from Abcam (Cambridge, MA). Cells were grown to 95% confluency in 35-mm dishes and lysed with lysis buffer (50 mM Tris, pH 7.5, 1% Triton X-100, 50 mM NaF, 200 μM Na₃VO₄, and 10 μg/ml pepstatin and leupeptin) for 20 min at 4°C. Lysates were centrifuged at 10,000 revolutions/min for 5 min. Proteins were separated using SDS-PAGE containing 15–30 μg protein on 7.5% polyacrylamide gels. Samples were transferred to an Immobilon-P membrane (Millipore, Bedford, MA) at 15 volts for 30 min. Membranes were blocked in 10% nonfat dehydrated milk in PBS with 0.1% Tween 20 (PBS-T) for 1 h at room temperature and then incubated with primary antibodies diluted in 10% nonfat dehydrated milk in PBS-T overnight at 4°C. Membranes were washed three times for 10 min each with PBS-T, incubated with the respective secondary antibodies conjugated to horseradish peroxidase (1:3,000 dilution) in PBS-T, washed again with PBS-T, and visualized using SuperSignal chemiluminescent substrate (Pierce, Rockford, IL) and the Chemidoc Imaging System (Bio-Rad, Hercules, CA). Quantification of protein expression was performed with Quality One software (Bio-Rad).

Filipin staining.

Cells were grown to 60–70% confluency on collagen-coated cover slips. Cells were rinsed three times with PBS, fixed with 4% paraformaldehyde for 30 min, and incubated with 0.05 mg/mL filipin (Sigma-Aldrich, St. Louis, MO) in PBS for 1 h on a shaker in the dark. Filipin was dissolved freshly in dimethyl sulfoxide before each experiment. Cells were rinsed and then mounted using SlowFade (Invitrogen) on slides. Cells were visualized in the ultraviolet range using a wide-field microscope with a ×40 oil objective on a Leica DM6000 upright microscope (×40 oil objective) with Improvision's Volocity software. The number of cells having or not having perinuclear cholesterol accumulation was counted, and quantification was determined by the ratio of cells with accumulation to total cells. Over 35 images were quantified for each condition.

Microtubule formation assay.

Cells were grown on collagen-coated cover slips to 75–80% confluency. Cells were placed on ice for 45–60 min for microtubule depolymerization. After the depolymerization period, prewarmed 37°C media (with carrier or drug) was added at designated time points. At the end of the time course, cells were rinsed with PBS and fixed with cold methanol immediately. Cells were rinsed with PBS and blocked with 5% goat serum for 30 min, incubated with α-tubulin antibodies (1:200 dilution) in 5% goat serum for 1 h, and incubated with Texas Red goat anti-rabbit IgG antibodies (1:4,000 dilution) in 5% goat serum in the dark. Cells were rinsed and mounted with SlowFade (Invitrogen) on slides. Cells were visualized in the appropriate range using a Leica DM6000 upright microscope (×40 oil objective) with Improvision's Volocity software. The number of cells having or not having an aster was counted, and quantification was determined by the ratio of cells with aster formation to total cells at various time points.

Immunostaining.

Rab7 (rabbit) antibodies were obtained from Abcam (Cambridge, MA). Cells were grown on collagen-coated cover slips to 75–80% confluency. Cells were rinsed and fixed with cold methanol for 10 min at −20°C. Cells were blocked in 5% goat serum for 30 min, incubated in 5% goat serum with Rab7 antibodies (1:150) for 1, and incubated with Texas Red goat anti-mouse IgG antibodies in 5% goat serum in the dark. Cells were rinsed and stained with1 µg/mL of 4',6-diamidino-2-phenylindole (DAPI) in PBS for 5 min in the dark. Cells were mounted with SlowFade (Invitrogen) on slides, and they were visualized in the appropriate range using a Leica DM6000 upright microscope (×40 oil objective) with Improvision's Volocity software.

Data analysis.

Using ImageJ software, perimeters were drawn around each nucleus, and partial perimeters highlighting the portion of the nucleus with endosomal accumulation were drawn on top of these. Measurements of the length of these segments were generated on the software and used to calculate the ratio of the covered nuclear circumference. A t test was performed to compare the ratios of perinuclear accumulation between resveratrol (RSV)-treated and mock-treated IB3 cells. Every cell with a nucleus fully visible in the frame of the image and that did not appear to be dividing was measured by this method and included in the analysis.

RESULTS

Resveratrol restores cholesterol processing in CF-model cells.

Two cultured cell models were used to initially test whether resveratrol treatment could restore cholesterol trafficking. We first used the immortalized CF cell line IB3 cells and corrected control S9 cells, since the cholesterol accumulation phenotype occurs in these cells (53). Resveratrol treatment (50 µM, 24 h) effectively reverses the perinuclear cholesterol accumulation, suggesting intracellular transport is restored (Fig. 1A). Cells with cholesterol accumulation decreased from 56.2 ± 8.9% to 21.3 ± 8.4% in IB3 cells when treated with RSV. In S9 cells, this number decreased from 36.9 ± 2.6% to 16.5 ± 3.6%. To test endosomal distribution directly, endosomal localization was determined by Rab7 immunostaining. In CF-model cells, it was shown that Rab7-positive organelles accumulate near the nucleus with a limited distribution, as shown previously (40). Resveratrol treatment leads to redistribution of the endosomes, consistent with improving intracellular transport (Fig. 1B). The mean percentage of Rab7-covered nuclear circumferences was 44.8% for mock-treated and 78.0% for RSV-treated IB3 cells. It was previously reported that inhibition of soluble adenylate cyclase with KH7 in WT epithelial cells mimics the CF phenotype of cholesterol accumulation and microtubule disruption (38). The ability of resveratrol to reverse the actions of KH7 on cholesterol transport was tested. Consistent with CF cell data, resveratrol (50 µM, 24 h) treatment completely restores cholesterol trafficking in KH7-treated cells (Fig. 1C). The percentage of cells with cholesterol accumulation was 10.4 ± 4.0% for cells treated with vehicle, 44.6 ± 12.5% for cells treated with KH7 only, 7.6 ± 5.2% for cells treated with RSV only, and 11.4 ± 5.0% for cells treated with both KH7 and RSV. These data demonstrate that resveratrol is an effective agent in reversing intracellular transport issues associated with CF-model cells.

Fig. 1.

Resveratrol restores cholesterol processing in cystic fibrosis (CF)-model cells. A: S9 and IB3 cells are treated with vehicle or with 50 µM of resveratrol (RSV) for 24 h and assessed for cholesterol distribution by filipin staining. Significance is determined by t test comparing vehicle vs. resveratrol treated for each cell line; *P < 0.001, n = 5 experiments. Representative images are shown for each condition. B: IB3 cells were treated with vehicle or 50 µM of RSV for 24 h, immunostained with Rab7 antibodies and DAPI. Images of cells were quantified as mentioned in Immunostaining. Over 85 cells of each condition from 5 separate experiments were measured. Significance was determined by t test to determine significance between groups; P < 0.001. Representative images are shown for each condition. C: 9/HTEo⁻ cells were treated with RSV (50 µM) in the absence and presence of KH7 (50 µM) for 24 h and assessed for cholesterol distribution by filipin staining as mentioned in methods. Significance was determined by ANOVA and Newman-Keuls multiple-comparison test to determine significance between groups. Significance is shown for groups compared with vehicle (NT), P < 0.05 (n = 5 experiments). Representative images are shown for each condition. Scale bars represent 50 µm. D: wild-type (WT) and cystic fibrosis transmembrane conductance regulator (Cftr)^−/− (CF) mouse nasal epithelium (MNE) cells were treated with RSV (50 µM) for 24 h and assessed for cholesterol distribution by filipin staining. Significance was determined by ANOVA and Newman-Keuls multiple-comparison test to determine significance between groups, n = 4 separate experiments. *P < 0.05 compared with WT untreated. **P < 0.05 compared with CF untreated.

It has been reported that RSV can lead to CFTR activation. To test whether CFTR function was necessary for the intracellular transport correction in CF cells, we tested the ability of RSV to improve cholesterol mobilization in primary mouse nasal epithelial (MNE) cells isolated from Cftr^−/− mice. RSV (50 µM) reduced cholesterol accumulation significantly in Cftr^−/− MNE, showing CFTR is not necessary for RSV efficacy (Fig. 1D). We also examined the impact of RSV treatment on F508del CFTR function using human nasal epithelial (HNE) cells from F508del patients. No enhancement of F508del CFTR activity is observed with RSV (Table 1). Together these data indicate that modulation of CFTR function is not a mechanism of RSV efficacy.

Table 1.

Effect of RSV on CFTR function in F508del/F508del HNE cells

	Isc		P Value
Treatment	DMSO	RSV (50 µM)
Baseline	81.9 ± 33.4	114.1 ± 2.6	NS
Amiloride (ΔIsc)	−1.6 ± 0.8	−2.5 ± 0.4	NS
Forskolin/vx-770 (ΔIsc)	−1.3 ± 2.3	−3.3 ± 1.1	NS
Inh172 (ΔIsc)	−3.5 ± 1.0	−4.0 ± 0.3	NS

Values are means ± SE; n = 5 experiments for each group. I_sc, short-circuit current. Units are µA/cm². Resveratrol (RSV) has no effect on cystic fibrosis transmembrane conductance regulator (CFTR) function in primary cystic fibrosis human nasal epithelial (HNE) cells homozygous for the F508del mutation. Cells were treated with 50 µM RSV for 24 h. There were no significant (NS) differences in resistance values after any treatment. Significance determined by t test.

To determine if the efficacy of RSV was dose dependent, a dose-response curve examining cholesterol mobilization (Fig. 2, A and C) was performed. Results show a clear dose response, with the first near-maximal effect dose being 50 µM. These studies are done with trans-RSV, so we also examined the efficacy of cis-RSV in mediating increased endosomal transport in IB3 cells. cis-RSV is statistically effective in correcting the phenotype in CF-model cells at higher doses, although less efficiently than trans-RSV (Fig. 2B).

Fig. 2.

Differential effect of trans-resveratrol (RSV) and cis-RSV on cholesterol distribution in cystic fibrosis (CF) IB3 cells. A: dose response of trans-RSV on cholesterol distribution. Significance was determined by ANOVA and Newman-Keuls multiple-comparison test to determine significance between groups, n = 5 separate experiments. *P < 0.05 compared with 0 µM RSV group. B: dose response of cis-resveratrol (RSV) on cholesterol distribution. Significance was determined by ANOVA and Newman-Keuls multiple-comparison test to determine significance between groups, n = 5 separate experiments. *P < 0.05 compared with 0 µM RSV group. C: representative images of cholesterol distribution in IB3 cells with varying doses of trans-RSV. Scale bars represent 50 µm.

Determination of resveratrol-mediated signaling mechanisms controlling intracellular transport.

The above data demonstrate that resveratrol treatment significantly improves both microtubule structure and intracellular transport in CF-model cells. We have previously demonstrated that AMPK activation corrects microtubule formation rates in primary CF epithelial cells whether stimulated by ibuprofen or 5-aminoimidazole-4-carboxamide ribonucleotide (AICAR; see Ref. 39). The effect of these compounds on microtubule regulation was completely blocked by the AMPK inhibitor compound C (40). Because resveratrol is capable of stimulating AMPK function, the role of this pathway in mediating the effects of resveratrol was tested. IB3 cells were treated with resveratrol (50 µM, 24 h) in the absence and presence of compound C (10 µM). Compound C had no effect on the ability of resveratrol to correct cholesterol distribution in IB3 cells (Fig. 3A). The percentage of cells with cholesterol accumulation was 70.2 ± 14.9% for cells treated with vehicle, 45.4 ± 8.6% for cells treated with RSV only, 74.2 ± 8.8% for cells treated with compound C only, and 46.1 ± 9.4% for cells treated with both compound C and RSV. Previous work has demonstrated that the direct cause of impaired intracellular transport in CF cells is changes to microtubule acetylation and reformation rates (37–39). We tested directly whether resveratrol could influence microtubule reformation rates in CF-model IB3 cells. Results demonstrate that resveratrol treatment does indeed increase the rate of reformation in IB3 cells. However, compound C also had no influence on resveratrol stimulation of microtubule formation rates in IB3 cells, further demonstrating a lack of AMPK signaling in mediating the effects of resveratrol (Fig. 3B).

Fig. 3.

Compound C (CC) could not reverse the effect of resveratrol on correction of cholesterol distribution or stimulation of microtubule formation rates in IB3 cells. A: IB3 cells were treated with resveratrol (RSV, 50 µM) in the absence and presence of compound C (10 µM) for 24 h and assessed for cholesterol distribution by filipin staining. Significance was determined by ANOVA and Newman-Keuls multiple-comparison test to determine significance between groups. Significance is shown for groups compared with vehicle (NT), *P < 0.05 (n = 5 experiments). Representative images are shown for each condition. B: IB3 cells were treated with RSV (50 µM) in the absence and presence of compound C (10 µM) for 24 h and analyzed by a microtubule formation assay. Images of cells were quantified as mentioned in Microtubule formation assay. The percentages of cells with asters in IB3 cells treated with vehicle, RSV, and RSV + compound C at various time points are shown. Microtubule formation rates were significantly increased in cells treated with RSV regardless of the presence of compound C at 2, 4, and 8 min (n = 5 experiments). Significance was determined by ANOVA and Newman-Keuls multiple-comparison test to determine significance between groups at each time point. Significance is shown for groups compared with vehicle (NT), *P < 0.05 (n = 5). Representative images are shown for each condition. Scale bars represent 50 µm.

Resveratrol stimulates microtubule formation in primary human nasal epithelial cells.

To determine if resveratrol had similar function in primary cells, F508del/F508del primary human nasal epithelial (HNE) cells were treated with resveratrol and assayed for microtubule reformation. Resveratrol significantly increases the rate of microtubule reformation, consistent with IB3 cell results (Fig. 4). These data demonstrate that the effect of resveratrol is not model specific and has direct impact on microtubule processes that we have previously shown to be related to intracellular transport (37–39).

Fig. 4.

Resveratrol stimulates microtubule formation in primary human nasal epithelial cells. Human nasal epithelial (HNE) cells were treated with resveratrol (RSV, 50 µM) for 24 h and analyzed by a microtubule formation assay. Images of cells were quantified as mentioned in Microtubule formation assay. The percentages of cells with asters in cells treated with vehicle or RSV at various time points are shown. Microtubule formation rates are significantly increased in cells treated with RSV at 1 and 2 min (n = 7 experiments). CF, cystic fibrosis; NT, no treatment (vehicle control). Significance is determined by t test for each time point between the two groups, *P < 0.01. Representative images of aster formation are shown for various time points. Scale bars represent 50 µm.

Phosphodiesterase inhibition likely not involved in resveratrol efficacy.

Resveratrol also acts as a cAMP phosphodiesterase (PDE) inhibitor (23, 35, 56), and we have demonstrated the ability of cAMP to modulate intracellular transport and microtubule dynamics in CF cells through EPAC1 (27, 38). Therefore, the ability of the PDE inhibitor isobutyl methylxanthine (IBMX, 200 µM, 24 h) to modulate cholesterol processing was examined. The percentages of cells with cholesterol accumulation were 44.4 ± 5.6% and 50.7 ± 7.5% for mock-treated and IBMX-treated S9 cells and 56.0 ± 5.8% and 59.3 ± 6.0% for mock-treated and IBMX-treated IB3 cells. There was no significant difference between mock-treated and IBMX-treated cells in either cell line. IBMX treatment had no impact on cholesterol accumulation, suggesting PDE inhibition is not a likely mechanism of resveratrol action (Fig. 5).

Fig. 5.

Isobutyl methylxanthine (IBMX) has on effect on cholesterol distribution in S9 and IB3 cells. S9 and IB3 cells were treated with vehicle or with 200 µM of IBMX for 24 h and assessed for cholesterol distribution by filipin staining. NT, no treatment (vehicle control). Significance was determined by t test to determine significance between groups. There was no significant difference (NS) between mock-treated and IBMX-treated cells in either cell line (n = 5 experiments). Representative images are shown for each condition. Scale bars represent 50 µm.

Sirtuin signaling does not mediate resveratrol effects.

The best-characterized role of resveratrol is as a direct activator of SIRT1 (13). Given this primary function, the role of sirtuins in mediating the effects of resveratrol on intracellular transport were examined. Initially, the ability of the sirtuin cofactor NAD⁺ was tested for its ability to restore cholesterol transport in the CF-model IB3 cells or control S9 cells. IB3 cells exhibit increased cholesterol accumulation compared with control S9 cells, consistent with previous studies, but NAD⁺ (5 mM, 24 h) had no significant impact in either cell line (Fig. 6A). The percentages of cells with cholesterol accumulation were 36.1 ± 16.8% and 32.2 ± 12.6% for mock-treated and NAD⁺-treated S9 cells and 56.5 ± 8.8% and 45.2 ± 10.6% for mock-treated and NAD⁺-treated IB3 cells. The ability of NAD⁺ to reverse cholesterol accumulation induced by the soluble adenylate cyclase inhibitor KH7 was also examined. Consistent with CF IB3 cell results, no impact of NAD⁺ was seen (Fig. 6B). The percentage of cells with cholesterol accumulation was 10.3 ± 7.0% for cells treated with vehicle, 40.1 ± 4.4% for cells treated with KH7 only, 8.3 ± 5.7% for cells treated with NAD⁺ only, and 33.4 ± 6.1% for cells treated with both KH7 and NAD⁺.

Fig. 6.

Sirtuin signaling does not mediate resveratrol effects. A: mock-treated or NAD⁺ (5 mM, 24 h)-treated S9 and IB3 cells were assessed for cholesterol distribution by filipin staining. Significance was determined by t test to determine significance between groups. There was no significant difference (NS) between mock-treated and NAD⁺-treated cells in either cell line (n = 5 experiments). Representative images are shown for each condition. B: 9/HTEo⁻ cells were treated with NAD⁺ (5 mM) in the absence and presence of KH7 (50 µM) for 24 h and assessed for cholesterol distribution by filipin staining. Significance was determined by ANOVA and Newman-Keuls multiple-comparison test to determine significance between groups. Significance is shown for groups compared with vehicle (NT), *P < 0.05 (n = 5 experiments). No significant effect of NAD⁺ treatment was found. Representative images are shown for each condition. Scale bars represent 50 µm.

To more directly test the role of sirtuins in mediating the effects of resveratrol, the ability of the SIRT1 inhibitor EX-527 and the SIRT1/3 inhibitor 4-bromo-resveratrol (4bRSV) to block resveratrol-mediated cholesterol trafficking in IB3 cells was tested. The percentage of cells with cholesterol accumulation was 69.1 ± 17.6%, 37.0 ± 11.8%, 68.1 ± 20.7%, and 43.2 ± 9.1% for cells treated with vehicle, RSV, EX-527, and RSV in addition to EX-527, respectively. The difference between cells treated with RSV and cells treated with both RSV and EX-527 was not significant. Similarly, the percentage of cells with cholesterol accumulation was 52.6 ± 5.1%, 26.4 ± 5.4%, 47.2 ± 13.3%, and 25.7 ± 7.7% for cells treated with vehicle, RSV, 4bRSV, and RSV in addition to 4bRSV, respectively. The difference between cells treated with RSV and cells treated with both RSV and 4bRSV was not significant. Neither compound had a significant impact on resveratrol efficacy, suggesting sirtuin activation is not a key factor in the mechanisms (Fig. 7, A and B).

Fig. 7.

EX-527 or 4-bromo-resveratrol (4bRSV) could not reverse the effect of resveratrol (RSV) on correcting cholesterol distribution. A: IB3 cells were treated with RSV (50 µM) in the absence and presence of EX-527 (1 µM) for 24 h and assessed for cholesterol distribution by filipin staining. Significance was determined by ANOVA and Newman-Keuls multiple-comparison test to determine significance between groups. Significance is shown for groups compared with vehicle (NT), *P < 0.05 (n = 5 experiments). Representative images are shown for each condition. B: IB3 cells were treated with RSV (50 µM) in the absence and presence of 4bRSV (15 µM) for 24 h and assessed for cholesterol distribution by filipin staining. Significance was determined by ANOVA and Newman-Keuls multiple-comparison test to determine significance between groups. Significance is shown for groups compared with vehicle (NT), *P < 0.05 (n = 5 experiments). Representative images are shown for each condition. Scale bars represent 50 µm.

Inhibition of peroxisome proliferator-activated receptor-γ signaling blocks effect of resveratrol on CF cells.

Resveratrol is also reportedly a selective activator of peroxisome proliferator-activated receptor-γ (PPARγ; see Ref. 33); thus, the role of the PPARγ pathway in mediating the effects of resveratrol on intracellular transport was examined. IB3 cells were treated with resveratrol (50 µM, 24 h) in the absence or presence of the PPARγ inhibitor GW-9662 (20 µM, 24 h). Unlike other interventions, GW-9662 successfully blocked the effect of resveratrol in reversing cholesterol accumulation in IB3 cells (Fig. 8). The percentage of cells with cholesterol accumulation was 76.7 ± 12.0%, 32.1 ± 15.7%, 79.8 ± 7.9%, and 72.8 ± 10.2% for cells treated with vehicle, RSV, GW-9662, and RSV in addition to GW-9662, respectively. These data demonstrate that resveratrol is acting at least in part through the PPARγ pathway to promote intracellular transport.

Fig. 8.

Peroxisome proliferator-activated receptor-γ (PPARγ) inhibition blocks the efficacy of resveratrol in restoring endosomal transport in cystic fibrosis (CF)-model cells. A: IB3 cells were treated with resveratrol (RSV, 50 µM) in the absence and presence of GW-9662 (20 µM) for 24 h and assessed for cholesterol distribution by filipin staining. Significance was determined by ANOVA and Newman-Keuls multiple-comparison test to determine significance between groups. Significance is shown for groups compared with vehicle (NT), *P < 0.05 (n = 5 experiments). RSV + GW-9662 is significantly different from RSV treatment alone, **P < 0.05. Representative images are shown for each condition. Scale bars represent 50 µm.

As a control, the ability of GW-9662 to inhibit ibuprofen-mediated correction of intracellular transport in IB3 cells was examined. GW-9662 treatment had no impact on ibuprofen-mediated correction of cholesterol transport in IB3 cells, demonstrating that resveratrol is acting through a mechanism independent of ibuprofen/AMPK and that GW-9662 is not initiating an effect of its own on intracellular transport (Fig. 9A). The percentage of cells with cholesterol accumulation was 66.4 ± 5.8%, 48.8 ± 3.1%, 69.4 ± 3.5%, and 48.1 ± 10.0% for cells treated with vehicle, ibuprofen, GW-9662, and both ibuprofen and GW-9662, respectively. Significant differences were found in cells treated with or without ibuprofen regardless of the presence of GW-9662, but GW-9662 had no impact on ibuprofen efficacy. Because the PPARγ inhibitor blocked the effects of resveratrol on intracellular transport, we tested whether PPARγ activation directly with the agonist pioglitazone could replicate the efficacy of resveratrol. CF-model IB3 cells were treated with pioglitazone (10 µM, 24 h) and assessed for cholesterol distribution by filipin staining. Pioglitazone treatment successfully reduces cholesterol distribution in IB3 cells (Fig. 9B). These data suggest that PPARγ signaling is necessary to mediate the effects of resveratrol treatment.

Fig. 9.

Peroxisome proliferator-activated receptor-γ (PPARγ) signaling is necessary but not sufficient to mediate the effects of resveratrol treatment. A: IB3 cells were treated with ibuprofen (IBU, 500 µM) in the absence and presence of GW-9662 (20 µM) for 24 h and assessed for cholesterol distribution by filipin staining. Significance was determined by ANOVA and Newman-Keuls multiple-comparison test to determine significance between groups. Significance is shown for groups compared with vehicle (NT), *P < 0.05 (n = 5). Representative images are shown for each condition. B: IB3 cells were treated with vehicle or with 10 µM of pioglitazone (PGZ) for 24 h and assessed for cholesterol by filipin staining. Significance was determined by t test to determine significance between groups; *P = 0.05 (n = 8 experiments). Representative images are shown for each condition. Scale bars represent 50 µm.

Resveratrol increases acetylated tubulin levels in CF cells.

Because PPARγ activation alone is not sufficient to explain the efficacy of resveratrol in CF cells, we examined another pathway relevant to intracellular transport control in CF cells. We have previously shown that microtubule acetylation is reduced in CF cells and tissues (37). Inhibiting the primarily cytosolic histone deacetylase family member HDAC6 restores tubulin acetylation and intracellular transport in CF cells (37). We tested whether resveratrol could impact tubulin acetylation. Resveratrol treatment significantly increases tubulin acetylation in CF-model IB3 cells (Fig. 10). These data are consistent with a previous report suggesting resveratrol could act as a pan-HDAC inhibitor and suggest that resveratrol’s effects on both PPARγ activation and HDAC6 inhibition likely mediate its effects on microtubule regulation and subsequently intracellular transport.

Fig. 10.

Resveratrol increases acetylated tubulin content in cystic fibrosis-model IB3 cells. Expressions of acetylated α-tubulin (Acetyl α-tub) in mock-treated and resveratrol (RSV, 50 µM, 24 h)-treated IB3 cells were analyzed via Western blot (n = 5 experiments). α-Tubulin was used for normalization of protein loading. NT, no treatment (vehicle control). Representative Western blot images were shown. Significance was determined by t test, *P < 0.01.

DISCUSSION

We have demonstrated previously that intracellular transport in CF epithelial cells is impaired with regard to endosomal movement (26). This impairment is marked by free cholesterol accumulation in late endosomes and is linked to growth and inflammatory signaling. We have identified a number of mechanisms that contribute to microtubule regulation in CF, including EPAC1 and HDAC6 activity (38). Correction of microtubule dynamics and associated intracellular transport functions is also achieved with AMPK activation with either direct AMPK agonists or the anti-inflammatory ibuprofen (39). The finding that ibuprofen so effectively impacted the microtubule pathway strongly suggested that targeting this pathway could be an important anti-inflammatory therapeutic target for CF to provide the benefits of ibuprofen therapy without the potential adverse effects.

The overall goal of this research is to identify accessible compounds that can stimulate intracellular transport in CF cells through correction of microtubule regulation to be evaluated for anti-inflammatory properties. In this study, it is hypothesized that resveratrol would be an effective agonist to stimulate microtubule elongation in CF cells and restore intracellular transport. Resveratrol is an attractive candidate, since it reportedly interacts with a number of pathways already identified to influence microtubule dynamics in CF. Resveratrol is a known agonist of AMPK signaling, and we have demonstrated that AMPK agonists potently stimulate microtubule formation in both cultured and primary CF cells, correcting endosomal movement (3, 7, 39). Resveratrol also acts as a phosphodiesterase inhibitor, increasing cAMP levels in various studies (52). We have demonstrated that cAMP activation of EPAC1 can regulate CF microtubule dynamics (38). The most direct impact of resveratrol is the stimulation of sirtuins, particularly SIRT1 (13). Both SIRT1 and SIRT3 have been shown to influence microtubules and offer a potentially novel mechanism for control of microtubules in CF (15, 30).

Resveratrol has been studied as a potential therapy in CF before, mostly in the context of CFTR activation. One study found resveratrol to be effective in stimulating CFTR function in acquired CFTR deficiency due to hypoxia (54). Other studies extend this property to CF disease-causing mutations. Cho et al. recently demonstrate that resveratrol augments the activation of G551D CFTR in the presence of modulators (4). One study even found evidence that resveratrol could help stimulate F508del CFTR function (5). The issue with these studies, as well as the data presented in this manuscript, is that higher doses of resveratrol in the 50- to 100-µM range are needed to achieve CFTR activation and other biological events. Resveratrol is not very bioavailable, with plasma concentrations only reaching ~2 µM (51). Indeed, Jai et al. have shown that physiologically achievable concentrations of resveratrol are insufficient to activate F508del CFTR (11).

Testing of resveratrol revealed that resveratrol is indeed a potent stimulator of microtubule formation in cultured models and primary CF epithelial cells. These microtubule effects result in more efficient intracellular transport as measured by cholesterol and Rab7-positive organelle distribution in CF-model IB3 cells. In this sense, the hypothesis tested positively. However, mechanistic analysis revealed that resveratrol exerts its effects through unexpected avenues. Data demonstrate that neither AMPK signaling, cAMP levels, nor sirtuin activation play a role in the actions of resveratrol on microtubule and intracellular transport. The lack of involvement of these predicted pathways required the identification of a more novel mechanism of action.

It was identified that resveratrol was exerting its influence on microtubules through two other mechanisms, PPARγ activation and likely HDAC6 inhibition. There is precedent for both of these mechanisms. Resveratrol activation of PPARγ has been reported previously, particularly as a mechanism for antiproliferation effects in tumor cells (1). Our data demonstrate that inhibition of PPARγ signaling effectively blocks the ability of resveratrol to restore intracellular transport in CF cells, effectively mimicked by direct PPARγ activation with pioglitazone. These findings are consistent with previous reports that PPARγ activity is reduced in CF epithelium and that its activation can have a number of beneficial effects, including improved immune function (9, 32, 33, 43, 46).

Another aspect of microtubule dysregulation in CF is reduced tubulin acetylation (37), and it has been reported that resveratrol can act as a pan-HDAC inhibitor (50). An examination of tubulin acetylation influence by resveratrol in CF cells does show a significant increase. These data are consistent with our previous findings that HDAC6 inhibition restores microtubule acetylation and intracellular transport in CF cells (37). Depletion of HDAC6 expression from CF mouse models also restores growth and inflammatory responses, demonstrating the importance of intervening in this pathway (34, 36).

The conclusion of the study is that resveratrol is a very effective modulator of microtubule regulation in cultured and primary CF epithelial cells, resulting in improved intracellular transport. The positive impact of resveratrol on microtubule formation rates and acetylation in CF cells is somewhat surprising. Several analogs of resveratrol have been developed as antitumor therapeutic candidates based on their ability to block tubulin polymerization (47). Although these analogs may be acting through different mechanisms than resveratrol itself, it is possible that resveratrol is acting on pathways that are specifically altered in CF, allowing for such dramatic improvement in microtubule formation. No adverse effects of resveratrol were observed on WT cell microtubules, but no further benefit was seen either. We also find that resveratrol treatment increases acetylated tubulin levels, which is consistent with resveratrol acting as a pan-HDAC inhibitor, including HDAC6, as mentioned. However, another report by Misawa et al. demonstrates that resveratrol treatment prevents acetylated tubulin accumulation in response to mitochondrial damage, preventing inflammasome activation (29).

Although further studies are necessary, the ability of resveratrol to reverse CF cellular phenotypes in a manner not only consistent, but more efficiently and with broader impact than ibuprofen, suggests that resveratrol could be an effective anti-inflammatory therapy. The anti-inflammatory properties of resveratrol are well established in multiple studies under a variety of conditions (10, 31, 41, 55). As noted above, a limitation of this study is that plasma levels of resveratrol cannot reach the 50-µM dose found to be necessary in these cellular data. However, understanding how resveratrol is acting in CF could lead to a new therapeutic avenue and to a better understanding of CF cell signaling.

GRANTS

This work is supported by Cystic Fibrosis Foundation (CFF) Grants KELLEY17G0 and CFF RDP R447-CR11.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

T.J.K. conceived and designed research; B.L. and D.A.C. performed experiments; B.L., D.A.C., and T.J.K. analyzed data; B.L. and T.J.K. interpreted results of experiments; B.L. and T.J.K. prepared figures; B.L. and T.J.K. drafted manuscript; B.L., D.A.C., and T.J.K. edited and revised manuscript; B.L., D.A.C., and T.J.K. approved final version of manuscript.