Curcumin stimulates cystic fibrosis transmembrane conductance regulator Cl- channel activity.

Compounds that enhance either the function or biosynthetic processing of the cystic fibrosis transmembrane conductance regulator (CFTR) Cl(-) channel may be of value in developing new treatments for cystic fibrosis (CF). Previous studies suggested that the herbal extract curcumin might affect the processing of a common CF mutant, CFTR-DeltaF508. Here, we tested the hypothesis that curcumin influences channel function. Curcumin increased CFTR channel activity in excised, inside-out membrane patches by reducing channel closed time and prolonging the time channels remained open. Stimulation was dose-dependent, reversible, and greater than that observed with genistein, another compound that stimulates CFTR. Curcumin-dependent stimulation required phosphorylated channels and the presence of ATP. We found that curcumin increased the activity of both wild-type and DeltaF508 channels. Adding curcumin also increased Cl(-) transport in differentiated non-CF airway epithelia but not in CF epithelia. These results suggest that curcumin may directly stimulate CFTR Cl(-) channels.

Cystic fibrosis (CF) 1 results from mutations in the gene encoding the CFTR Cl Ϫ channel (1). The most common CF mutation (⌬F508) causes defective protein folding, and as a result CFTR is targeted for degradation rather than progressing to the cell membrane (2). Other CF-associated mutations disrupt CFTR function by a variety of mechanisms, including some that prevent normal biosynthesis or impair CFTR Cl Ϫ channel function in the cell membrane. Knowledge of how mutations disrupt function has encouraged development of compounds that correct CFTR-⌬F508 processing and/or stimulate CFTR activity for those channels in the cell membrane (3). This effort has been spurred on by the suggestion that even a fraction of normal CFTR activity may be enough to prevent disease; studies in non-CF humans and in mice bearing targeted alterations of the CFTR gene suggest that that 5-10% of wild-type CFTR levels may be sufficient to prevent lung disease in humans and intestinal disease in mice (4 -7). Two additional discoveries support the feasibility of a pharmacologic approach. First, although the Cl Ϫ channel activity of CFTR-⌬F508 and several other CF-associated mutants is reduced, they do retain significant Cl Ϫ channel activity (2,8,9). Second, defective processing of CFTR-⌬F508 can be partially corrected by reducing the incubation temperature or adding chemical chaperones (8,9).
In pursuit of this strategy, a recent study found that curcumin partially corrected the processing defect of CFTR-⌬F508 in heterologous cells (10). Moreover, administering curcumin to mice homozygous for the ⌬F508 mutation corrected the characteristic defect in voltage across the nasal epithelium, whereas nasal voltage was unaltered in null mice. Curcumin administration also improved weight gain in CFTR-⌬F508 animals. However, three more recent studies did not find an effect of curcumin on CFTR-⌬F508 processing (11)(12)(13).
Given the potential value of agents that modify CFTR function and/or processing and given differences in the reported effects of curcumin, we asked if curcumin might stimulate CFTR Cl Ϫ channel activity. Earlier studies had suggested that a very small amount of CFTR-⌬F508 may reach the cell surface in mouse intestinal and airway epithelia (14,15). Thus, an agent that stimulates CFTR-⌬F508 activity might increase Cl Ϫ transport independent of alterations in mutant protein processing. Likewise, if an agent produced even a small correction in CFTR-⌬F508 processing, then increasing its channel activity would further augment the physiologic consequences. Our interest in curcumin stimulation of CFTR was also piqued by the earlier suggestion (10, 12) that curcumin exhibits structural similarities to other compounds that may bind directly to CFTR and stimulate its activity; examples include genistein, apigenin, benzo[c]quinoliziniums, and benzimidazolones (16 -20). Therefore, we tested the hypothesis that curcumin stimulates the activity of wild-type and ⌬F508 CFTR Cl Ϫ channels.

EXPERIMENTAL PROCEDURES
Cells and Expression Systems-For patch clamp studies, wild-type and mutant CFTR were transiently expressed in HeLa cells using a hybrid vaccinia virus system as described previously (21). For studies of CFTR phosphorylation, COS-7 cells were electroporated with pcDNA3-CFTR (22).
Cultures of human airway epithelia were obtained from non-CF and CF bronchus (⌬F508/⌬F508) and cultured at the air-liquid interface as described previously (23). Epithelia were used at least 14 days after seeding when they were well differentiated with a surface consisting of ciliated cells, goblet cells, and other non-ciliated cells (23). These differentiated epithelia retain the functional properties of airway epithelia including transepithelial electrolyte transport.
CFTR Phosphorylation-48 h after transfection, COS cells were washed three times with ice-cold phosphate-buffered saline and solubilized for 1 h at 4°C with ice-cold lysis buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Nonidet P-40 and proteinase inhibitors) (22). Soluble proteins were centrifuged for 15 min at 15,000 ϫ g at 4°C. Proteins in soluble supernatant were immunoprecipitated overnight at 4°C with anti-CFTR antibodies (24-1 from R&D Systems, Minneapolis, MN, and MM13-4 and M3A7 from Upstate Technologies, Charlottesville, VA). The antibodyprotein complex was precipitated with protein A-agarose (Pierce). Immunoprecipitated CFTR was washed three times with ice-cold lysis buffer containing detergent and once with lysis buffer without detergent. Pellets were then phosphorylated with 40 nM catalytic subunits of PKA (Promega, Madison, WI) and [␥-32 P]ATP (6000 Ci/mM, PerkinElmer Life Sciences) plus the indicated concentrations of curcumin or the Me 2 SO vehicle control. Phosphorylation was stopped by dilution into 100ϫ volume of lysis buffer and cAMP-dependent protein kinase peptide inhibitor (PKI) (10 M) (Promega). Proteins were separated by SDS-PAGE, stained, destained, dried, and exposed to phosphor screens (Amersham Biosciences). Radioactivity was quantitated using the ImageQuant program (Amersham Biosciences).
An Axopatch 200A amplifier (Axon Instruments, Inc., Union City, CA) was used for voltage clamping and current recording, and the pCLAMP software package (version 8.1, Axon Instruments, Inc.) was used for data acquisition and analysis. Recordings were low pass Bessel-filtered at 10 kHz and stored using a digital tape recorder. Replayed recordings were low pass-filtered at 100 Hz using an 8-pole Bessel filter (Model 900, Frequency Devices, Inc., Haverhill, MA) and digitized at 250 Hz for macropatch recordings. Single-channel recordings were low pass-filtered at 700 Hz and digitized at 7 kHz. Singlechannel analysis was performed as described previously (27,28) with a burst delimiter of 20 ms. Events Յ 15 ms in duration were ignored.
Reagents-Curcumin was obtained from Sigma (catalog no. C7727, lot no. 093K0795) and Fluka Chemika (Seelze, Germany, product no. 28260, lot no. 453335/144203356); we obtained similar results with both. Curcumin was dissolved in Me 2 SO prior to use and added directly to experiments without intermediate dilution in aqueous buffers. As we began our patch clamp experiments with curcumin, we noticed that when a curcumin solution was placed in a plastic syringe reservoir and gravity fed through a 50-cm length of 0.58-mm diameter polyethylene tubing (Clay Adams® Intramedic), the color of the solution emerging was less yellow than the starting material. Absorbance spectra of the samples showed a peak at 418 nm, consistent with curcumin, and after perfusion through the tubing more than 70% of the peak disappeared. Therefore, we prepared curcumin and added it directly to the cells or chambers in all studies. We were not able to study concentrations Ͼ 50 M because of poor solubility. For prolonged exposure, curcumin was dissolved in Me 2 SO, diluted into culture medium at 1:1000, and added to the submucosal surface for 3, 6, or 18 h.
Data-Data are shown as means Ϯ S.E. unless otherwise indicated. p values Ͻ0.05 were considered statistically significant.

Curcumin Stimulates CFTR in Excised Membrane
Patches-To learn whether curcumin could stimulate CFTR, we added it to the cytosolic surface of excised, inside-out membrane patches containing many CFTR channels. We added curcumin dissolved in Me 2 SO directly to the cytosolic bath solution, because when we applied it through a perfusion system it adsorbed to the tubing, and we observed no effect on current (see "Experimental Procedures"). Curcumin stimulated a large, reversible increase in Cl Ϫ channel activity (Fig. 1A). Stimulation required the presence of ATP and only occurred in channels that had been phosphorylated by PKA (not shown).
As the curcumin dose increased, the current response increased (Fig. 1B). Previous studies showed that highly phosphorylated channels (those studied in the presence of PKA) have an altered response to the channel activator genistein and to ATP concentration (29,30). Therefore, we also examined the effect of curcumin on highly phosphorylated channels (i.e. in the presence of PKA). Adding curcumin after removing PKA stimulated current with an apparent EC 50 of 9.1 Ϯ 2.4 M. Curcumin appeared more potent when added in the presence of PKA, EC 50 2.2 Ϯ 1.1 M. However in highly phosphorylated channels, 50 M curcumin stimulated less current than 10 M (Fig. 1B).
Curcumin Inhibits PKA Activity-Finding that 50 M curcumin stimulated less current than 10 M curcumin in highly phosphorylated channels raised the possibility that curcumin might inhibit channel activity. Such behavior was reported for genistein, which stimulated CFTR currents at low micromolar concentrations and inhibited currents at concentrations above 20 M (29, 31). Alternatively, curcumin might alter PKA activity. To test for a direct effect of curcumin on PKA, we asked whether it inhibited PKA phosphorylation of CFTR. We found that as the curcumin concentration increased, CFTR phosphorylation fell (Fig. 2). Curcumin has also been reported to inhibit protein kinase C activity (32). The reduced PKA activity may explain, in part, the decreasing current in response to high curcumin concentrations in the presence of PKA. However, we cannot exclude the possibility that high curcumin concentrations have a small direct inhibitory effect on CFTR.
Curcumin Stimulates CFTR-⌬F508 -To learn whether curcumin stimulates CFTR-⌬F508, we cultured cells at 27°C to allow mutant channels to reach the cell surface (8,9). When we then excised membrane patches, we found that even though the amount of current was less than with wild-type channels, 10 M curcumin stimulated current (Fig. 3).
Curcumin Increases Channel Opening and Slows Channel Closing-To understand how curcumin increases current, we examined the response of single channels. Fig. 4A shows an example. Curcumin augmented activity by elevating P o (Fig.  4B). The increase resulted from both a prolongation of burst duration and a reduction in interburst interval. Curcumin did not alter single-channel current amplitude.
Curcumin and Genistein Do Not Have Additive Effects-As noted previously (10, 12), curcumin has some structural similarities to genistein and other compounds that stimulate CFTR activity. Therefore, we compared the effects of genistein and curcumin and found that 10 M curcumin generated a greater current increase than 10 M genistein (Fig. 5). To further explore the mechanism by which curcumin stimulates CFTR, we added curcumin and genistein together. Instead of additive effects, genistein reduced the magnitude of curcumin-stimulated current. These results suggest that the two compounds might have related binding sites.
Curcumin Acutely Increases Cl Ϫ Current in Differentiated Non-CF Airway Epithelia-To learn whether curcumin stimulates epithelial Cl Ϫ transport, we measured the transepithelial Cl Ϫ current across well differentiated human airway epithelia. As observed previously, cAMP agonists increased Cl Ϫ current in non-CF but not CF epithelia (Fig. 6, A and B). We found that curcumin slowly increased Cl Ϫ current in non-CF epithelia, but it failed to alter current in CF epithelia. These results are consistent with curcumin stimulation of CFTR. Ten M curcumin had little effect (Fig. 6B), even though this concentration stimulated substantial activity in isolated CFTR Cl Ϫ channels (Fig. 1B). The difference may relate to different access to CFTR channels in the two systems, different levels of CFTR phosphorylation, variable adsorption, or other factors.
Our patch clamp studies showed that curcumin only stimulated activity after channels had been phosphorylated. In the epithelial studies, CFTR has a basal level of phosphorylation and activity (33,34) that would allow curcumin to stimulate transepithelial Cl Ϫ current. We found that curcumin stimulated less current than cAMP agonists, and adding curcumin and cAMP agonists together generated no more current than cAMP agonists alone (Fig. 6C). We obtained similar results measuring bumetanide-sensitive current (Fig. 6C), which provides an assay of total CFTR-dependent transepithelial Cl Ϫ transport (23).
We also tested the effect of more prolonged curcumin addition. Curcumin had little effect on Cl Ϫ current in non-CF epithelia (Fig. 7). CF epithelia (⌬F508/⌬F508) developed no bumetanide-sensitive current even after incubation with a range of curcumin concentrations and exposure for 3 and 6 h (not shown) or 18 h (Fig. 7). DISCUSSION In these studies, we focused on the acute response to curcumin and found that it stimulates CFTR Cl Ϫ channel activity. Stimulation was dose-dependent and reversible and occurred with both wild-type and ⌬F508 channels. Curcumin also stimulated CFTR activity in differentiated human airway epithelia.
Finding that curcumin stimulated CFTR in excised, insideout membrane patches suggests that it might interact directly with CFTR. Although our data do not allow us to determine the molecular mechanism, they do provide some initial sugges-tions. We found that curcumin only stimulated phosphorylated CFTR, and its potency increased when applied to highly phosphorylated channels. Moreover, stimulation occurred at curcumin concentrations that partially inhibit PKA phosphorylation of CFTR. These results suggest that the primary mechanism probably does not directly involve the R domain (35,36). The lack of effect on single-channel current amplitude also suggests that major changes in the anion pore were probably not responsible. Nevertheless, we cannot exclude the possibility that curcumin induced structural changes in the R domain or membrane-spanning domains that somehow influenced channel function. However, because of its effects on gating, we favor the hypothesis that curcumin altered nucleotide binding domain (NBD)-dependent regulation. Curcumin reduced the duration of the closed state; this gating step is determined by nucleotide concentration, nucleotide structure, and NBD amino acid sequence (35,36). Curcumin also prolonged the burst duration, a gating step influenced by NBD mutations and by interventions that alter NBD enzymatic activity. In speculating about the mechanism of stimulation, we can also compare curcumin to other compounds that stimulate CFTR. Earlier studies noted that the biphenolic compound curcumin shares some structural features with flavones (apigenin), isoflavones (genistein), benzimidazolones (NS004), and benzo[c]quinolizinium compounds that increase channel activity (16 -20, 37). In addition, the monophenolic compound capsaicin was reported to stimulate CFTR (38). Like curcumin, genistein and capsaicin also prolonged the burst duration and reduced the interburst interval (29,38,39). Although the mechanism(s) by which these compounds stimulate is not known with certainty, some work suggests that genistein interacts with the CFTR NBDs to alter gating (31, 40 -42). Our finding that genistein and curcumin do not have additive effects and, further, that genistein attenuated the larger current increase generated by curcumin suggest that the two compounds may share common mechanisms, perhaps involving the NBDs. It will be interesting to learn more about how these various agents stimulate CFTR and whether their functional effects depend on shared structural features. A better understanding of the relationship between structure and mechanism of stimulation might suggest ways to improve activity.
Previous studies examined the effects of administering curcumin to cells and mice for hours or days with a goal of testing its effect on the CFTR-⌬F508 processing defect (10 -13). Those studies used different methods and animals and came to different conclusions about whether curcumin facilitated CFTR-⌬F508 transit through the biosynthetic pathway to emerge at the cell surface and rescue the CF Cl Ϫ channel defect. We did not attempt to reproduce those earlier studies. However, we observed no effect of prolonged curcumin exposure on Cl Ϫ transport by differentiated CF airway epithelia. There are many potential explanations for our failure to detect a response, including the very low level of CFTR-⌬F508 expression in these epithelia. Nevertheless, our data suggest that a curcumin-induced increase in Cl Ϫ channel activity might have contributed, at least in part, to the effects of curcumin described previously (10).
Our results highlight some potential difficulties in assessing the effect of curcumin. Curcumin adsorption to plastic and solubility were concerns. Curcumin stimulated the CFTR Cl Ϫ channel. Curcumin inhibited PKA activity. Like genistein, curcumin might inhibit current at high concentrations. Also previous studies suggested curcumin may enhance processing of CFTR-⌬F508 (10). Given the variety of these effects, which might be compounded by methodological differences between laboratories, it is perhaps not surprising to see differences in reports about the activities of curcumin. FIG. 7. Effect of prolonged curcumin addition to differentiated airway epithelia. Data are change in bumetanide-inhibited transepithelial current in non-CF and CF (⌬F508/⌬F508) airways after 18 h of incubation with curcumin in basolateral medium. Prior to the addition of bumetanide, epithelia were treated with amiloride, 4,4Ј-diisothiocyanotostilbene-2,2Ј-disulfonic acid, and cAMP agonists. n ϭ 4 for each non-CF condition; n ϭ 3 for each CF condition. We obtained similar results with 10 M (n ϭ 3) or 50 M (n ϭ 15) curcumin added to epithelia obtained from three different CF (⌬F508/⌬F508) individuals. In addition, curcumin added for 3-6 h did not induce a bumetanide-sensitive current (not shown).