Additive effect of multiple pharmacological chaperones on maturation of CFTR processing mutants

Abstract

The most common cause of CF (cystic fibrosis) is the deletion of Phe⁵⁰⁸ (ΔF508) in the CFTR [CF TM (transmembrane) conductance regulator] chloride channel. One major problem with ΔF508 CFTR is that the protein is defective in folding so that little mature protein is delivered to the cell surface. Expression of ΔF508 CFTR in the presence of small molecules known as correctors or pharmacological chaperones can increase the level of mature protein. Unfortunately, the efficiency of corrector-induced maturation of ΔF508 CFTR is probably too low to have therapeutic value and approaches are needed to increase maturation efficiency. We postulated that expression of ΔF508 CFTR in the presence of multiple correctors that bound to different sites may have an additive effect on maturation. In support of this mechanism, we found that expression of P-glycoprotein (CFTR's sister protein) processing mutants in the presence of two compounds that bind to different sites (rhodamine B and Hoechst 33342) had an additive effect on maturation. Therefore we tested whether expression of ΔF508 CFTR in the presence of combinations of three different classes of corrector molecules would increase its maturation efficiency. It was found that the combination of the quinazoline VRT-325 together with the thiazole corr-2b or bisaminomethylbithiazole corr-4a doubled the steady-state maturation efficiency of ΔF508 CFTR (approx. 40% of total CFTR was mature protein) compared with expression in the presence of a single compound. The additive effect of the correctors on ΔF508 CFTR maturation suggests that they directly interact at different sites of the protein.

INTRODUCTION

CFTR [CF (cystic fibrosis) TM (transmembrane) conductance regulator] is a cAMP-regulated chloride channel located in the apical membrane of polarized epithelia, where it plays a key role in regulating salt and water homoeostasis [1]. CF is a lethal inherited disease caused by mutations in CFTR that inhibit folding of the protein or its channel activity [2]. Defects in CFTR biosynthesis or channel activity cause accumulation of thick dehydrated mucus in the airways that ultimately leads to deterioration and eventual failure of the lungs due to chronic lung infections.

Deletion of Phe⁵⁰⁸ (ΔF508) is the most common CF mutation as it is found on at least one chromosome in 90% of affected individuals [3]. The ΔF508 protein is defective in folding so it fails to reach the cell surface because it accumulates in the ER (endoplasmic reticulum) and is rapidly degraded [4,5].

Expression of ΔF508 CFTR at low temperature [6] or in the presence of a chemical chaperone such as glycerol [7] can increase the level of functional ΔF508 CFTR at the plasma membrane. The low temperatures (27–30 °C) or high levels of glycerol (10%, v/v) required for ΔF508 CFTR rescue are not feasible for use with CF patients. Short-chain fatty acids such as butyrate or sodium 4-phenylbutyrate promote ΔF508 CFTR maturation by increasing CFTR expression [8,9]. Although 4-phenylbutyrate has been approved by the FDA (Food and Drug Administration) to treat urea-cycle disorders, it has limitations such as rapid metabolism, resulting in the requirement for high oral doses (19 g/day) and limited efficacy [10,11].

Misprocessed CFTR is retained in the ER by molecular chaperones. SERCA (sarcoplasmic/endoplasmic-reticulum Ca²⁺-ATPase) inhibitors, such as curcumin [12] and thapsigargin [13], can deplete the calcium stores in the ER and are postulated to disrupt chaperone–ΔF508 CFTR interactions to allow the protein to be trafficked to the plasma membrane. The use of SERCA inhibitors to promote maturation of ΔF508 CFTR has drawbacks as it may only function in specific cell lines or strains of mice [14,15], and it may have harmful effects if used in patients due to disruption of multiple metabolic pathways.

Studies on processing mutants of CFTR's sister protein, the multidrug P-gp (P-glycoprotein), showed that specific pharmacological chaperones (also known as correctors) could be used to correct folding defects in processing mutants [16,17]. Therefore high-throughput screening of chemical libraries was performed to identify potential pharmacological chaperones for ΔF508 CFTR. A number of potential pharmacological chaperones were identified such as corr-2b, corr-3a, corr-4a and corr-4b [18], VRT-325 [17,19,20] and VRT-532 [20]. Some compounds such as corr-4a, corr-2b and VRT-532 may directly interact with CFTR, since they showed specificity for rescue of ΔF508 CFTR [20]. VRT-325 may be able to bind directly to both P-gp and CFTR [21]. A problem with the best correctors identified to date, such as VRT-325 [22] or corr-4a [18], is that the level of correction of ΔF508 CFTR may be too low to have therapeutic value. A potential strategy to increase the efficiency of CFTR maturation may be to use multiple correctors. In the present study, we tested whether expression of P-gp or CFTR processing mutants in the presence of combinations of corrector compounds would increase the efficiency of maturation.

MATERIALS AND METHODS

Construction and expression of mutants

The construction of ΔF508 and H1085R CFTR cDNAs and insertion into the pcDNA3 vector (Invitrogen Canada, Burlington, ON, Canada) was described previously [23]. The CFTR cDNAs coding for ΔNBD2 (where NBD2 is nucleotide-binding domain 2), H1085R CFTR (residues 1–1196) as well as P-gp cDNAs for mutant P709G P-gp and P-gp truncation mutants ΔNBD2 P-gp (residues 1–1023), TMD1+2 {TMD1 [N-terminal TMD (TM domain) containing TM segments 1–6] and TMD2 (C-terminal TMD containing TM segments 7–12)}, P-gp (residues 1–379 plus 681–1025) and TMD1 P-gp (residues 1–379) were modified to contain the epitope for monoclonal antibody A52 at the C-terminal ends and subcloned into the mammalian expression vector pMT21 as described previously [21]. Transient expression of CFTR or P-gp was carried out by transfection of HEK-293 cells (human embryonic kidney cells).

Effect of compounds on misprocessed mutants

HEK-293 cells were transfected with cDNAs of mutant P-gp or CFTR. On the next day, the medium was replaced with DMEM (Dulbecco's modified Eagle's medium) with 10% (v/v) calf serum [or 2% (v/v) for CFTR correctors] containing various concentrations of compounds and the cells were incubated at 37 °C under 5% CO₂. The cells were harvested after 24 h and solubilized with 2× SDS sample buffer [125 mM Tris/HCl, pH 6.8, 20% (v/v) glycerol, 4% (w/v) SDS and 4% (v/v) 2-mercaptoethanol] containing 50 mM EDTA. Whole cell samples were then subjected to SDS/PAGE (6% acrylamide gels) and immunoblot analysis with rabbit polyclonal antibody against CFTR or with the mouse monoclonal antibody A52. The amount of mature CFTR or P-gp relative to total (mature plus immature CFTR or P-gp) was quantified by scanning the gel lanes followed by analysis with the NIH (National Institutes of Health) Image program (http://rsb.info.nih.gov/nih-image/). Results are means for three independent experiments.

Treatment with endoglycosidases

Cells expressing ΔF508 CFTR were solubilized with 2× SDS sample buffer containing 50 mM EDTA. For treatment with endo H (endoglycosidase H), a 1/10 vol. of 0.5 M sodium citrate (pH 5.5) was added to cell extract followed by addition of 20000 units/ml of endo H (New England Biolabs, Mississauga, ON, Canada). The sample was treated for 15 min at 20 °C. For treatment with PNGase F (peptide N-glycosidase F), a 1/10 vol. of 0.5 M sodium phosphate (pH 7.5) and 1/10 vol. of 10% (v/v) Nonidet P40 were added to the cell extract followed by addition of 10000 units/ml PNGase F (New England Biolabs). Samples were incubated for 15 min at 37 °C and then subjected to immunoblot analysis as described above.

Cell-surface labelling

BHK (baby-hamster kidney) cells stably expressing mutant ΔF508 CFTR were grown in DMEM containing 2% (v/v) calf serum for 48 h at 37 °C in the absence or presence of 3 μM VRT-325, 10 μM corr-2b or 3 μM VRT-325 plus 10 μM corr-2b. The cells were washed four times with PBSCM [PBS (pH 7.4) containing 0.1 mM CaCl₂ and 1 mM MgCl₂] and then treated in the dark with PBSCM buffer containing 10 mM NaIO₄ for 30 min at 20 °C. The cells were then washed four times with PBSCM buffer and treated with sodium acetate buffer (100 mM sodium acetate buffer, pH 5.5, 0.1 mM CaCl₂ and 1 mM MgCl₂) containing 2 mM biotin-LC-hydrazide (Pierce, Rockford, IL, U.S.A.) for 30 min at 20 °C. The cells were then washed twice with sodium acetate buffer and solubilized with Tris-buffered saline (100 mM Tris/HCl, pH 7.4, and 150 mM NaCl) containing 1% (w/v) Triton X-100 and protease inhibitors (Cocktail Set III; Calbiochem). CFTR was immunoprecipitated with monoclonal antibody M3A7 and subjected to SDS/PAGE on 6% gels, and biotinylated CFTR was detected with streptavidin-conjugated horseradish peroxidase and enhanced chemiluminescence.

RESULTS

One useful approach to prevent protein misfolding and trafficking defects of CFTR processing mutants is to carry out synthesis in the presence of specific compounds called correctors or pharmacological chaperones [18,21,22,24]. A problem with the correctors identified to date is that their efficiency in promoting maturation of mutants such as ΔF508 CFTR is quite low. The usefulness of correctors as potential treatments for CF would be increased if methods can be identified to enhance their ability to promote maturation of CFTR mutants. One potential strategy that may enhance maturation of CFTR processing mutants would be to utilize combinations of correctors that bind to different sites in CFTR. Unfortunately, the locations of corrector-binding sites in CFTR have not yet been identified.

A useful tool that would enable us to test the prediction that binding of multiple correctors to different binding sites would have an additive effect on maturation is P-gp. P-gp is a useful model system to aid understanding of maturation of CFTR processing mutants for several reasons. First, both CFTR and P-gp are predicted to have similar structures (see Figure 1) as they are both members of the ABC (ATP-binding-cassette transporter) family of transporters. Secondly, processing mutations of CFTR introduced at homologous positions in P-gp also inhibit maturation of the protein. For example, deletion of Tyr⁴⁹⁰ (equivalent to ΔF508 in CFTR) inhibits maturation of P-gp [25]. Thirdly, processing mutants of P-gp can be rescued by carrying out expression in the presence of drug substrates [16]. Finally, P-gp has up to four distinct drug-binding sites located in the TMDs [26] and the protein can simultaneously bind multiple drug substrates [27–29].

Figure 1. Models of P-gp (A) and CFTR (B).

The cylinders represent TM segments, the glycosylation sites are represented by branched lines and R represents the regulatory domain. The filled circles represent the locations of processing mutations examined in the present study. The structure of ΔNBD2 and TMD1+2 P-gp and ΔNBD2 CFTR are also shown.

P-gp contains distinct binding sites for the drug substrates Hoechst 33342 (H site) and rhodamine 123 (R site). These sites were reported to interact in a positively co-operative manner [9]. In addition, rhodamine B can promote maturation of P-gp processing mutants [30]. We wished to now test whether rhodamine B and Hoechst 33342 would act co-operatively to promote maturation of P-gp processing mutants. Mutations in the linker region between the two halves of P-gp (Figure 1A) such as P709G inhibit the maturation of the protein. We modified the cDNA for P-gp P709G to encode an epitope for monoclonal antibody A52 to allow us to distinguish the expressed protein from low levels of endogenous P-gp that is found in HEK-293 cells [31]. The first step was to examine the concentration dependence of rhodamine B and Hoechst 33342 on maturation of mutant P709G. Figure 2(A) shows that P-gp mutant P709G is expressed as an immature protein with an apparent mass of 150 kDa in the absence of drug substrate. P-gp contains three glycosylation sites in the extracellular loop between TM segments 1 and 2 (Figure 1A). The glycosylation sites are core-glycosylated in the ER to yield a 150 kDa protein that is sensitive to digestion with endo H [32]. The carbohydrate of the mature protein is modified in the Golgi to yield a 170 kDa P-gp that is resistant to endo H but sensitive to PNGase F. The presence of mature P709G P-gp with an apparent mass of 170 kDa could be detected in the presence of 30 μM rhodamine B (Figure 2A). Higher concentrations of rhodamine B were toxic to the cells. In contrast, little maturation was observed when mutant P709G was expressed in the presence of 0.1–3 μM Hoechst 33342 (Figure 2B). Concentrations of Hoechst 33342 higher than 1 μM were toxic to the cells as they began to lift off the tissue culture plate. This result suggests that rhodamine B induced some maturation of mutant P709G, while Hoechst 33342 was less effective. HEK-293 cells transiently expressing P709G P-gp were then studied with or without 25 μM rhodamine B, 1 μM Hoechst 33342 or a combination of the two compounds for 24 h. These concentrations of drug were selected because they were the highest combination of concentrations that did not cause the cells to lift off the tissue culture plate. Whole cell SDS extracts of the treated cells were then subjected to immunoblot analysis with monoclonal A52 antibody. We compared the efficiency of rescue by quantifying the amount of mature P-gp relative to the total amount of P-gp. The amount of mature P-gp increased from less than 10% in the absence of drug to 21 and 16% when it was expressed in the presence of rhodamine B and Hoechst 33342 respectively. When the mutant was expressed in the presence of both compounds, the yield of mature protein reached 52% of total P-gp (Figure 2C). Therefore expression of mutant P709G in the presence of both rhodamine B and Hoechst 33342 had an additive effect on maturation.

Figure 2. Rescue of a P-gp processing mutant with a mutation in the linker region (P709G).

HEK-293 cells were transiently transfected with P-gp P709G and then incubated for 24 h in the presence of 0–30 μM rhodamine B (A) or 0–3 μM Hoechst 33342 (B). (C) HEK-293 cells were transiently transfected with P-gp P709G and then incubated for 24 h in the absence (–) or presence (+) of 25 μM rhodamine B, 1 μM Hoechst 33342 or combination of 25 μM rhodamine B and 1 μM Hoechst 33342. Whole cell extracts were subjected to immunoblot analysis with monoclonal antibody A52. The positions of mature and immature P-gps are shown.

Previously, we had shown that the TMDs alone are sufficient to mediate drug binding for compounds such as cyclosporin A, verapamil, vinblastine and capsaicin [33]. To test whether the combination of rhodamine B and Hoechst 33342 promote maturation through an interaction with the TMDs, we examined P-gp truncation mutants lacking NBD2 (ΔNBD2 P-gp) or both NBDs (TMD1+2), because both proteins are normally expressed as immature proteins (see Figures 3A and 3B, lanes 1). Rhodamine B (25 μM) or Hoechst 33342 (1 μM) were very inefficient in promoting maturation of ΔNBD2 P-gp when used separately, as little increase in mature protein was observed. When a combination of the two compounds was used however, a significant increase in the mature product was observed that gave a maturation efficiency of 33% (Figure 3A). The construct lacking both NBDs, TMD1+2 P-gp, was then tested. Mature TMD1+2 P-gp was only observed upon expression in the presence of both rhodamine B and Hoechst 33342 (Figure 3B). In addition, we performed a study on a P-gp ‘quarter-molecule’ (TMD1) [34], but found that neither each compound nor a combination of the two compounds induced maturation of the protein (results not shown). This result suggests that rhodamine B and Hoechst 33342 binding requires the presence of at least the two TMDs to promote maturation of P-gp.

Figure 3. Effect of Hoechst 33342 and rhodamine B on maturation of P-gp truncation mutants.

HEK-293 cells were transfected with A52-tagged ΔNBD2 P-gp (A) or TMD1+2 (B) followed by a 24 h incubation in the absence (–) or presence (+) of 25 μM rhodamine B, 1 μM Hoechst 33342 or combination of 25 μM rhodamine B and 1 μM Hoechst 33342 respectively. Whole cell extracts were subjected to immunoblot analysis with monoclonal antibody A52. The positions of mature and immature P-gps are indicated.

The P-gp results support the prediction that expression of processing mutants in the presence of multiple compounds could enhance maturation of CFTR processing mutants. Although CFTR is structurally similar to P-gp, CFTR also contains a regulatory domain (Figure 1B). Several CFTR correctors such as VRT-325, corr-4a and corr-2b were identified by high-throughput screening approaches. The quinazoline derivative, VRT-325, can promote maturation of both P-gp and CFTR processing mutants [22]. corr-4a is a bisaminomethylbithiazole compound that promotes maturation of ΔF508 CFTR but shows little rescue of a mutant dopamine receptor [18]. The thiazole derivative corr-2b is also specific as it also promotes maturation of ΔF508 CFTR but shows little or no rescue of a P-gp processing mutant [20]. Therefore VRT-325, corr-4a and corr-2b were selected for study because they are the most efficient correctors of ΔF508 CFTR maturation and they are structurally distinct so may bind to CFTR at different sites.

First, we tested the correctors on a CFTR processing mutant (H1085R) with a mutation in the fourth cytoplasmic loop (CL4) connecting TM segments 10 and 11 in the CO₂H half of the molecule, where a relatively large number of clinically relevant processing mutations are found [35]. The H1085R CFTR mutant was selected for use in our initial studies because expression of its cDNA in HEK-293 cells yields a small amount of mature protein (approx. 10% of total CFTR) [23]. Therefore the use of mutant H1085R in the presence of combinations of correctors would enable us to determine if the compounds increase or decrease the efficiency of maturation. CFTR processing mutant H1085R was transiently expressed in HEK-293 cells and incubated in the presence or absence of 3 μM VRT-325, 10 μM corr-2b and 10 μM corr-4a, combination of VRT-325 and corr-2b, combination of VRT-325 and corr-4a, or combination of corr-2b and corr-4a for 24 h. Whole cell extracts were subjected to immunoblot analysis. Maturation of CFTR can be monitored by a difference in mobility on SDS/PAGE gels between the core-glycosylated immature protein and mature protein containing complex carbohydrate due to N-glycosylation at two sites located between TM segments 7 and 8 (Figure 1B). The yield of mature CFTR with VRT-325, corr-2b or corr-4a alone was 25, 28 and 15% respectively (Figure 4). Expression of mutant H1085R CFTR in the presence of a combination of VRT-325 and corr-2b caused the largest increase in maturation efficiency as 48% of total CFTR was present as mature protein. The combination of VRT-325 and corr-4a increased the maturation efficiency to 39%. The combination of corr-2b and corr-4a yielded a level of mature protein that was similar to that observed with corr-2b alone.

Figure 4. Effect of correctors on maturation of CFTR processing mutant H1085R.

CFTR processing mutant H1085R was expressed for 24 h in the absence (–) or presence (+) of 3 μM VRT-325, 10 μM corr-2b, 10 μM corr-4a, combination of 3 μM VRT-325 and 10 μM corr-2b, combination of 3 μM VRT-325 and 10 μM corr-4a, or combination of 10 μM corr-2b and 10 μM corr-4a respectively. Whole cell extracts were then subjected to immunoblot analysis with a rabbit polyclonal antibody against CFTR. The positions of mature and immature CFTRs are indicated.

It has been postulated that correctors promote maturation of CFTR processing mutants by promoting interactions between the TMDs or between the NBDs to form the characteristic nucleotide-sandwich dimer [18]. To determine if interactions between the NBDs were required for the enhanced maturation of mutant H1085R in the presence of VRT-325 together with corr-2b or corr-4a, we utilized an H1085R mutant that lacked NBD2 (ΔNDB2 H1085R CFTR) (Figure 1B). The ΔNDB2 H1085R CFTR mutant was transfected into HEK-293 cells and expressed in the presence of VRT-325, corr-2b, corr-4a, VRT-325 plus corr-2b, VRT-325 plus corr-4a or no correctors for 24 h. Whole cell extracts were then subjected to immunoblot analysis. No mature CFTR protein was observed in the absence of correctors and only small amounts were observed when the mutant was expressed in the presence of a single corrector. It was found that combination of VRT-325 and corr-2b was the most effective combination to promote maturation (55% of total CFTR was mature protein) (Figure 5). Some increase was also observed with the combination of VRT-325 plus corr-4a (37%).

Figure 5. Effect of correctors on maturation of CFTR truncation mutant ΔNBD2 H1085R.

HEK-293 cells were transfected with ΔNBD2 H1085R followed by a 24 h incubation in the absence (–) or presence (+) of 3 μM VRT-325, 10 μM corr-2b, 10 μM corr-4a, combination of 3 μM VRT-325 and 10 μM corr-2b, or combination of 3 μM VRT-325 and 10 μM corr-4a respectively. Whole cell extracts were subjected to immunoblot analysis with monoclonal antibody A52. The positions of mature and immature CFTRs are indicated.

Next we examined the most frequent mutant in CF, ΔF508 CFTR. Since maturation of ΔF508 CFTR is very inefficient, we utilized a stable line of BHK cells [22] rather than transient expression in HEK-293 cells. The use of the BHK stable cell line enabled us to carry out longer incubations in the presence of correctors and avoid problems of the short expression windows that could occur during transient transfection of HEK-293 cells. BHK cells stably expressing ΔF508 CFTR were incubated for 72 h in the presence of 3 μM VRT-325, 10 μM corr-2b, 10 μM corr-4a or pairs of these correctors. Whole cell extracts were then subjected to immunoblot analysis. Little mature CFTR was observed (less than 5% of total CFTR protein) when the cells were expressed in the absence of correctors (Figure 6A). The immunoblot showed only immature protein that was sensitive to endo H (Figure 6B). Expression in the presence of correctors VRT-325, corr-2b or corr-4a alone increased the steady-state levels of mature CFTR to 21, 8 and 9% of total CFTR respectively. Expression of the cells in the presence of VRT-325 together with corr-2b or corr-4a was particularly effective in increasing the levels of mature CFTR to 44 and 32% respectively. We confirmed that the slow migrating species was mature CFTR as the protein was sensitive to PNGase F but resistant to endo H (Figure 6B).

Figure 6. Effect of correctors on maturation of CFTR processing mutant ΔF508.

(A) Mutant ΔF508 CFTR in a BHK stable cell line was expressed for 72 h in the absence (None) or presence of 3 μM VRT-325 (325), 10 μM corr-2b (2b), 10 μM corr-4a (4a), combination of 3 μM VRT-325 and 10 μM corr-2b (325/2b), combination of 10 μM corr-2b and 10 μM corr-4a (2b/4a), or combination of 3 μM VRT-325 and 10 μM corr-4a (325/4a) respectively. (B) Samples from cells treated with no correctors (None) or treated with a combination of VRT-325 and corr-2b (325/2b) were treated with endo H (H), PNGase F (F) or no endoglycosidases (–). Whole cell extracts were then subjected to immunoblot analysis with a rabbit polyclonal antibody against CFTR. (C) Mutant ΔF508 CFTR in a BHK stable cell line was expressed in the absence (None) or presence of 3 μM VRT-325 (325), 10 μM corr-2b (2b) or 3 μM VRT-325 plus 10 μM corr-2b (325/2b) for 48 h. The cells were then subjected to cell-surface labelling with biotin-LC-hydrazide. CFTR was immunoprecipitated with monoclonal antibody M3A7 and subjected to SDS/PAGE. Biotinylated CFTR was detected with streptavidin conjugated to horseradish peroxidase and enhanced chemiluminescence. The positions of mature, immature and unglycosylated (Unglycos) CFTRs are indicated.

Cell-surface labelling was then performed to test if expression of mutant ΔF508 CFTR in the presence of corr-2b plus VRT-325 would increase the amount of CFTR at the cell surface relative to that with only corr-2b or VRT-325. BHK cells stably expressing mutant ΔF508 CFTR were incubated in the absence or presence of 3 μM VRT-325, 10 μM corr-2b or 3 μM VRT-325 plus 10 μM corr-2b. The cells were treated with sodium periodate to oxidize the carbohydrate groups and then treated with biotin-LC-hydrazide. The cells were solubilized with Triton X-100 and CFTR recovered by immunoprecipitation with monoclonal antibody M3A7. The immunoprecipitated CFTR was subjected to SDS/PAGE and biotinylated CFTR was detected with streptavidin-conjugated horseradish peroxidase and enhanced chemiluminescence. Figure 6(C) shows that expression of mutant ΔF508 CFTR in the presence of corr-2b plus VRT-325 increased the amount of CFTR at the cell surface by 2–4-fold compared with expression in the presence of only corr-2b or VRT-325. These results confirm that the mature CFTR rescued by the correctors is at the cell surface.

DISCUSSION

A potential treatment for patients with CF would be to promote folding of ΔF508 CFTR, since this mutation is found on at least one allele of approx. 90% of patients [3]. Expression of ΔF508 CFTR at low temperature (27 °C) [6] or in the presence of chemical chaperones such as glycerol [7] or trimethylamine N-oxide [36] can promote maturation and delivery of ΔF508 CFTR to the plasma membrane to yield functional channels. Unfortunately, it is not feasible to treat patients at 27 °C and chemical chaperones are too toxic to be used at the concentrations that induce rescue of ΔF508 CFTR. For example, treatment of mice expressing ΔF508 CFTR with sufficient trimethylamine N-oxide to rescue the protein resulted in a 50% mortality rate [37]. Short-chain fatty acids, such as sodium 4-phenylbutyrate, promote ΔF508 CFTR maturation by increasing CFTR mRNA synthesis and expression of Hsc70 (heat-shock cognate protein 70) and Hsp70 (heat-shock protein 70) that are believed to be involved in the protein maturation [8,9]. In a pilot clinical study, oral administration of sodium 4-phenylbutyrate to patients homozygous for ΔF508 CFTR partially restored nasal epithelial CFTR function [10]. The authors concluded that the levels of CFTR activity were very low however and that improved CFTR rescue approaches would be needed to treat CF patients [10,37].

Another potential approach to promote delivery of processing mutants to the cell surface is to alter interactions of the proteins with molecular chaperones. Misprocessed mutants of CFTR and P-gp are retained in the ER with molecular chaperones [38–40]. SERCA inhibitors such as curcumin [12] or thapsigargin [13] are postulated to deplete calcium stores in the ER and disrupt chaperone interactions with misprocessed CFTR to allow the protein to be trafficked to the plasma membrane. There are problems with the chaperone-disruption approach however as its ability to promote maturation of ΔF508 CFTR is highly variable [14,15] and the approach may have unpredictable side effects because many cellular metabolic pathways could be altered.

High-throughput screening of chemical libraries has recently identified various high-affinity correctors that can promote maturation of misprocessed CFTR mutants [18,24]. The mechanisms of these correctors are unknown. VRT-325 is not specific for CFTR as it can correct folding defects in processing mutants of CFTR [24], P-gp [22], as well as a processing mutant (G601S) of the hERG (human ether-a-go-go) cardiac potassium channel [24]. Evidence suggests, however, that VRT-325 may be able to bind directly to P-gp or CFTR. First, VRT-325 was shown to inhibit the ability of P-gp to confer resistance on cytotoxic compounds as well as its verapamil-stimulated ATPase activity [22]. Secondly, VRT-325 blocks disulfide cross-linking between cysteine residues located in TM6(L339C) and TM7(F728C) of P-gp [21]. Finally, VRT-325 appeared to interact with P-gp or CFTR during synthesis to modulate the topology of the protein [21]. Corrector compounds corr-4a and corr-2b may also directly interact with CFTR processing mutants because they show specificity in rescue of processing mutants. For example, corr-4a promotes maturation of ΔF508 CFTR but not a processing mutant of the dopamine 4 receptor [18]. corr-2b promotes rescue of CFTR processing mutants but not P-gp processing mutants [20]. Classes of compounds such as VRT-325, corr-2b and corr-4a are of interest because they are relatively more effective in promoting maturation of ΔF508 CFTR compared with compounds that are predicted to interfere with molecular chaperone interactions or transcomplementation approaches [20]. Although correctors such as corr-4a appear to be the most effective types of compounds to induce maturation of ΔF508 CFTR, it has been concluded that the maturation efficiency is still too low to be useful for treatment of CF patients [18].

An approach that may improve the usefulness of correctors to promote maturation of ΔF508 CFTR would be to utilize combinations of compounds that bind to different sites in CFTR. Interaction of compounds at different sites in CFTR may have an additive effect on maturation. P-gp was a useful model system to test a multiple binding site approach because it has been demonstrated that compounds such as rhodamine B and Hoechst 33342 bind to distinct sites [41,42]. We found that rhodamine B or Hoechst 33342 were relatively inefficient in promoting maturation of a P-gp processing mutant when they were used alone (see Figure 2). The efficiency of maturation of the P-gp processing mutant could be improved by approx. 2-fold when it was expressed in the presence of both compounds. The increase in the number of drug–P-gp contacts may be the reason for the increase in maturation efficiency when both compounds were used. Hydrogen bond interactions between P-gp and drug molecules appear to be important for promoting maturation of P-gp processing mutants [32].

Enhanced maturation of CFTR processing mutants was also observed when expression was carried out in the presence of pairs of corrector molecules (Figures 4 and 6). For ΔF508 CFTR, expression in the presence of VRT-325 and corr-2b or corr-4a increased the efficiency of maturation by approx. 2-fold compared with the use of VRT-325 alone or approx. 4-fold greater than corr-2b or corr-4a alone (Figure 6). It appears that the correctors affect packing of the TM segments because they can promote maturation of CFTR truncation mutants lacking one or both NBDs [21]. The additive effect of the multiple correctors on CFTR maturation suggests that they may interact at different sites in CFTR. One mechanism of the correctors may be to act as ‘pharmacological anchors’ to retain the unstable TM segments in the lipid bilayer [21]. Formation of a nucleotide-sandwich dimer does not seem to be essential for maturation of CFTR [43] or promotion of folding with multiple correctors (the present study).

In conclusion, the use of multiple correctors may be a useful technique for improving the maturation efficiency of CFTR processing mutants. Different classes of correctors acting at different sites may help to stabilize CFTR processing mutants as they are synthesized in the ER. Future studies will be needed to map the locations of the corrector-binding sites in CFTR.