Selective Inhibition of Endoplasmic Reticulum-associated Degradation Rescues ΔF508-Cystic Fibrosis Transmembrane Regulator and Suppresses Interleukin-8 Levels

Endoplasmic reticulum (ER)-associated degradation (ERAD) is the major quality control pathway of the cell. The most common disease-causing protein folding mutation, ΔF508-cystic fibrosis transmembrane regulator (CFTR), is destroyed by ERAD to cause cystic fibrosis (CF). p97/valosin-containing protein (VCP) physically interacts with gp78/autocrine motility factor receptor to couple ubiquitination, retrotranslocation, and proteasome degradation of misfolded proteins. We show here that p97/VCP and gp78 form complexes with CFTR during translocation from the ER for degradation by the cytosolic proteasome. Interference in the VCP-CFTR complex promoted accumulation of immature CFTR in the ER and partial rescue of functional chloride channels to the cell surface. Moreover, under these conditions, interleukin-8 (IL8), the expression of which is regulated by the proteasome, was reduced. Inhibition of the proteasome with bortezomib (PS-341/Velcade) also rescued CFTR, but with less efficiency, and suppressed NFκB-mediated IL8 activation. The inhibition of the major stress-inducible transcription factor CHOP (CCAAT/enhancer-binding protein homologous protein)/GADD153 together with bortezomib was most effective in repressing NFκB-mediated IL8 activation compared with interference of VCP, MLN-273 (proteasome inhibitor), or 4-phenylbutyrate (histone deacetylase inhibitor). Immunoprecipitation of ΔF508-CFTR from primary CF bronchial epithelial cells confirmed the interaction with VCP and associated chaperones in CF. We conclude that VCP is an integral component of ERAD and cellular stress pathways induced by the unfolded protein response and may be central to the efficacy of CF drugs that target the ubiquitin-proteasome pathway.

the ER lumen and are ultimately destroyed by the cytoplasmic ubiquitin-proteasome system. The ubiquitin-proteasome system plays a pivotal role in cell homeostasis and is vital in regulating various cellular processes. In normal cells, nearly all proteins are continuously degraded and replaced by de novo synthesis. The spatial separation between substrate selection and degradation in ERAD requires substrate transport from the ER to the cytoplasm by a process termed dislocation, recently reviewed by Meusser et al. (2).
The most common disease-causing protein folding mutation is deletion of phenylalanine at position 508 of the cystic fibrosis transmembrane regulator (⌬F508-CFTR), which results in a temperature-sensitive folding defect and premature degradation by ERAD (3). The absence of CFTR at the airway epithelial cell surface disrupts luminal hydration and is associated with an exaggerated immune response (4). Functional CFTR can be restored by growth at lower temperature (25-27°C) or incubation with chemical chaperones, which rescues ⌬F508-CFTR from ERAD (5).
Mammalian p97/valosin-containing protein (VCP) and its yeast counterpart, Cdc48, participate in retrotranslocation of misfolded proteins from the ER for degradation by the cytosolic proteasomes (6). p97/VCP and its cofactors (Ufd1, Npl4, and p47) interact with misfolded ubiquitinated substrates to dislodge them from the ER to the cytosol for proteasome degradation (7). We have reported recently that p97/VCP physically interacts with gp78 to couple ubiquitination, retro translocation, and proteasome degradation (8). Originally identified as the tumor autocrine motility factor receptor, gp78 is a multimembrane-spanning protein with RING finger-type ubiquitin-protein ligase activity exposed to the cytoplasmic ER surface. gp78/autocrine motility factor receptor is thus distinguished from CHIP (C terminus of Hsc70-interacting protein) and Parkin ubiquitin-protein ligases and closely resembles the yeast ERAD ubiquitin-protein isopeptide ligase Hrd1p/Der3p. In yeast, Hrd1p/Der3p and Cdc48 are required for CFTR degradation (9). We hypothesize that selective inhibition of ERAD not only rescues ⌬F508-CFTR, but also suppresses interleukin (IL)-8 levels, the major inflammatory cytokine in cystic fibrosis (CF) airways.

EXPERIMENTAL PROCEDURES
Human Subjects-CF (homozygous for ⌬F508-CFTR; n ϭ 3, 7-9 years old) and non-CF (n ϭ 3, 8 -17 years old) subjects who were undergoing fiberoptic bronchoscopy and bronchoalveolar lavage (BAL) for a clinical indication were invited to participate in a Johns Hopkins Institutional Review Board-and General Clinical Research Center-approved protocol. All three children with CF were chronically infected with one or more bacteria. The non-CF subjects were children with airway pathology: 1) a case of focal bronchiectasis from an earlier Mycoplasma pneumoniae infection (the BAL fluid was culture-positive for Branhamella catarrhalis, and the cytopathology contained neutrophils and lipid-laden macrophages); 2) a case of chronic lung disease and yellow nail syndrome (culture-positive for Hemophilus influenza and associated with neutrophils and lipid-laden macrophages upon BAL); and 3) a case of psychogenic cough (culture-negative and associated with lipid-laden macrophages upon BAL). Except for the case of psychogenic cough, there was a mild inflammatory phenotype. Parents signed informed consent on behalf of their children. Bronchial brushings were obtained after BAL to minimize contamination with mucus and neutrophils. Brushes were directly immersed in cell culture medium on ice for later transport to the laboratory.
Fluorescence Microscopy-After 48 h of transfection, cells were washed three times with Hanks' balanced salt solution containing calcium and magnesium (Invitrogen) and loaded with 1 M ER-Tracker TM Blue-White DPX dye (Invitrogen). Plates were incubated at 37°C and 5% CO 2 for 2 h and examined under a Zeiss Axiovert 135-N fluorescence microscope. A Quantix 1401 charge-coupled device camera and IPLab software Version 3.5 with appropriate filter settings for GFP (fluorescein isothiocyanate) or ER-Tracker Blue (4Ј,6-diamidino-2-phenylindole) were used to capture the images.
Immunoprecipitation and Immunoblotting-Cells were lysed directly on plates using M-PER protein lysis buffer (Pierce) containing protease inhibitor mixture (Roche Applied Science) after three washes with ice-cold phosphate-buffered saline. CF bronchial brushings were transferred to lysis buffer. For immunoprecipitation, total protein extracts (500 g/ml for cell culture and 50 g/ml for bronchial brushings) were incubated with 50 l of protein A/G-agarose beads (Santa Cruz Biotechnology, Inc.) for 3 h at 4°C. After preclearing, 5 g of respective primary antibody or preimmune serum (negative control) was added to each tube. After 1 h, protein A/Gagarose beads (50 l) were added to each tube, and tubes were incubated overnight at 4°C. Beads were washed once with lysis buffer (20 mM Tris-HCl (pH 7.6), 150 mM NaCl, 0.5% Triton X-100, and 10 M phenylmethylsulfonyl fluoride), followed with two washes with phosphate-buffered saline. The beads were suspended in Laemmli sample buffer (30 l) containing ␤-mercaptoethanol, vortexed for 1 min, resolved by 4 -10% SDS-PAGE, and transferred to a 0.4-m pore size nitrocellulose membrane. Proteins were detected using the respective primary antibodies. For pulsechase experiments, CFTR immunoprecipitate was eluted with sample buffer and separated on a 4 -8% SDS-polyacrylamide gel, dried for 2 h, and processed for autoradiography.
IL8 Cytokine Enzyme-linked Immunosorbent Assay-IB3-1 cells were transfected with the ⌬VCP, VCPQQ, ⌬gp78, or gp78⌬C construct and treated with 5 M bortezomib or 5 mM 4-PBA. At 24 h post-transfection, the cells were induced with 1 ng/ml IL1␤. After 48 h of transfection, supernatants were collected, and IL8 levels were measured by solid-phase enzyme amplified sensitivity immunoassay (BIOSOURCE) as specified by the manufacturer. Standards and high and low cytokine controls were included. The plates were read at 450 nm on a 96-well microplate reader (Molecular Devices Corp.) using SoftMax Pro software. The mean blank reading was subtracted from each sample and control reading. The amount of substrate turnover was determined calorimetrically by measuring the absorbance proportional to IL8 concentration. A standard curve was plotted, and an IL8 concentration in each sample was determined by interpolation from the standard curve. The data represent the means Ϯ S.D. of three independent experiments.
NoShift NFB Binding Assay-IB3-1 cells transfected (24 h) with ⌬VCP, ⌬F508-CFTR, or ⌬CHOP were treated with 5 M bortezomib (6 h), 5 M MLN-273 (6 h), or 5 mM 4-PBA (24 h). Cells were induced with IL1␤ (1 ng/ml) for 12 h. After 24 h of transfection, nuclear extracts (11) were collected using a NucBuster protein extraction kit (Novagen, San Diego, CA), and the shift in bound NFB was measured using a NoShift NFB binding assay kit (Novagen) as specified by the manufacturer. The signal from nuclear extracts was compared with the negative control (minus the extract). The specificity of protein binding was established using NFB-specific and mutant competitors. The plates were read at 450 nm on a 96-well microplate reader using SoftMax Pro software. The absorbance readings were used to determine the levels of bound NFB. The data represent the means Ϯ S.D. of three independent experiments.
N-(Ethoxycarbonylmethyl)-6-methoxyquinolinium Bromide (MQAE) Assay-CFTR-mediated chloride transport activity in IB3-1 cells transfected with the ⌬VCP, VCPQQ, ⌬gp78, or gp78⌬C construct and treated with bortezomib (1, 5, or 10 M) was assayed using the halidesensitive dye MQAE. Fluorescence of MQAE is quenched by halides. Cells were grown on glass coverslips in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum. Cells were loaded with the dye by hypotonic shock at 37°C for 40 min in Opti-MEM/water (1:1) containing 10 mM MQAE, washed, and then mounted in a perfusion chamber for fluorescence measurements. Fluorescence was measured using a Zeiss inverted microscope coupled to a charge-coupled device camera. Excitation and emission wavelengths were 350 and 460 nm, respectively. The stage and objective lenses were maintained at 37°C. The medium was kept flowing through the perfusion chamber at 1 ml/min using a peristaltic pump. Both 10 M forskolin and 10 M genistein were used to induce Cl Ϫ transport in IB3-1 cells. CFTR was inhibited with either 5 M CFTR inhibitor-172 (12) or myristoylated protein kinase A inhibitor 14 -22 amide (Calbiochem) to reduce cAMP-mediated phosphorylation. Chloride efflux was calculated using Equations 1-3.
Statistical Analysis-All data are represented as the means Ϯ S.D. of three experiments. One-way analysis of variance with Dunnett's planned comparison was run for each sample versus the control. A p value Ͻ0.05 was considered to have statistical significance.

RESULTS
p97/VCP Participates in ERAD of ⌬F508-CFTR-To test the hypothesis that VCP is involved in ⌬F508-CFTR degradation, we first estimated VCP protein levels in freshly isolated bronchial epithelial cells from ⌬F508 CF and non-CF subjects. As expected, VCP levels were up-regulated in all ⌬F508 CF subjects compared with non-CF controls (Fig. 1A). Moreover, the difference between CF and non-CF bronchial expression was striking. We next hypothesized that transient knockdown of either VCP or gp78 would rescue ⌬F508-CFTR from ERAD in CF bronchial epithelial cells. To reduce the expression of these proteins, we constructed short hairpin/small interfering RNA. The efficiency of VCP shRNA (⌬VCP)-mediated inhibition of VCP protein levels in IB3-1 cells (⌬F508/W1282X CF bronchial epithelial line) (13) was evaluated by immunoblotting of whole cell lysates (Fig. 1B, upper panel). To localize ⌬F508-CFTR signal, the IB3-1 cells were transfected with ⌬F508-CFTR-GFP and cotransfected with ⌬VCP or a plasmid control. Inhibition of VCP was associated with accumulation of ⌬F508-CFTR-GFP in the ER compared with the control (Fig. 1B, lower panels). Accumulation of ⌬F508-CFTR in the ER was confirmed by co-localization of the GFP signal with the signal from ER-Tracker Blue-White DPX. Furthermore, spreading of the GFP signal beyond the ER tracker dye compartment suggested that ⌬F508-CFTR-GFP escaped from the ER in ⌬VCP IB3-1 cells (Fig. 1C).
p97/VCP Inhibition Rescues ⌬F508-CFTR from ERAD-In IB3-1 cells, the ER-retained ⌬F508-CFTR protein is core-glycosylated (160-kDa immature B form), whereas growth at 26°C allows the ⌬F508-CFTR protein to transit the biosynthetic pathway and to acquire complex glycosylation (180-kDa mature C form) (5). Moreover, the ⌬F508-CFTR band C form has a faster turnover rate from the plasma membrane (14). To assess the effects of ⌬VCP on the stability of ⌬F508-CFTR bands B and C, IB3-1 cells were first transiently transfected with ⌬F508-CFTR-GFP for 48 h and then metabolically labeled with [ 35 S]methionine. ⌬F508-CFTR was immunoprecipitated with anti-CFTR antibody 169 (15) after the indicated chase time (Fig.   FIGURE 1. Inhibition of p97/VCP expression results in accumulation of CFTR in the ER. A, total protein extract (10 g) from freshly isolated bronchial epithelial cells obtained upon bronchoscopy from ⌬F508 CF subjects and non-CF controls was immunoblotted for VCP and ␤-actin. VCP is overexpressed in CF airways. B, IB3-1 cells were transfected with ⌬F508-CFTR-GFP and cotransfected with VCP shRNA (⌬VCP). The ⌬VCP efficiency is shown in the upper panel, in which the VCP protein level was drastically decreased by ⌬VCP (ϩ) compared with the control lane (Ϫ) (20 g in each lane). In a parallel experiment in the lower panels, ⌬F508-CFTR-GFP accumulated in a perinuclear location when VCP was decreased. C, IB3-1 cells transfected with ⌬F508-CFTR-GFP and/or ⌬VCP were loaded with ER-Tracker Blue-White DPX. The localization of ER-Tracker Blue dye and ⌬F508-CFTR-GFP is shown in upper and middle panels, respectively. The lower panel shows the bright-field image of the area indicating the normal morphology of cells after VCP inhibition. Co-localization of ER-Tracker Blue dye and ⌬F508-CFTR-GFP demonstrated accumulation of ⌬F508-CFTR in the ER. VCP inhibition augmented ⌬F508-CFTR accumulation in the ER, and part of it appeared to escape from the ER in ⌬VCP IB3-1 cells. Scale bar ϭ 20 m.
2A, Control). Cotransfection with ⌬VCP ( Fig. 2A) resulted in stabilization of both B and C forms. By comparison, cotransfection with gp78 small interfering RNA (⌬gp78) produced only a transient rescue of the mature C form with a peak at 1 h ( Fig. 2A), suggesting that VCP (not gp78) is central to extraction of CFTR from the ER for ERAD. S9 cells (IB3-1 cells corrected with wt-CFTR) were similarly transfected with additional wt-CFTR and ⌬VCP. We observed an efficient maturation and stability of band B to band C (Fig. 2B), confirming that wt-CFTR and ⌬F508-CFTR are extracted and degraded through the same VCP-mediated pathway.
We considered the proteasomal compartment as a potential therapeutic target for CF not only to test the role of VCP in proteasome- mediated degradation of CFTR, but also because both VCP and proteasome inhibitors are known to stabilize IB, the inhibitor of the NFB-mediated inflammatory response (16,17). We used bortezomib (6-h treatment) at 10 M, which leads to partial inhibition of protease activity in IB3-1 cells. Phase microscopic examination revealed no difference in morphology of treated and untreated monolayers (data not shown). To study the effect of bortezomib on another epithelial cell type homozygous for ⌬F508-CFTR, we transfected CFTE cells (homozygous for ⌬F508-CFTR) (18) with ⌬F508-CFTR-GFP. After 42 h of transfection, cells were induced with bortezomib (10 M) for 6 h. We observed a significant increase in the accumulation and stabilization of the B form with 10 M bortezomib compared with the untreated control (Fig. 2C). S9 cells transfected with wt-CFTR and exposed to bortezomib similarly produced substantial amounts of mature band C (Fig. 2B).
To study the effect of inhibiting components of ubiquitination and proteasome degradation on ERAD, we first compared the relative basal levels of the immature B and mature C forms of ⌬F508-CFTR in IB3-1 cells by metabolic labeling (t ϭ 2 h) (Fig. 2D, lane 1). We observed no changes in the levels of the B and C forms of ⌬F508-CFTR upon transfecting IB3-1 cells with empty vectors (pCIneo/pcDNA3/pSM2) compared with the control (lanes 2-4). In contrast, both VCP and gp78 inhibition resulted in accumulation of the B form compared with the control, whereas the C form showed a more significant increase with VCP inhibition compared with a minimal increase with gp78 inhibition (lanes 7 and 8). Because VCP physically interacts with gp78 (8), it is expected that the gp78-VCP interaction may represent one way of coupling ubiquitination with retrotranslocation and degradation of ⌬F508-CFTR. To evaluate the effect of this interaction on ERAD of ⌬F508-CFTR, we used a gp78⌬C construct deficient in the VCP-binding domain; no change in the levels of the B and C forms of ⌬F508-CFTR compared with the control was observed (lane 5), suggesting that this gp78-VCP interaction is required for ERAD of ⌬F508-CFTR. Moreover, the VCPQQ construct (deficient in AAA ATPase domains D1 and D2) partially rescued the B and C forms of ⌬F508-CFTR (lane 9), indicating The level of the C form was lower after transfection with the VCPQQ construct compared with complete VCP gene inhibition (⌬VCP) probably due to residual VCP AAA ATPase domain activity. Interference with Hsp70 expression (lane 6) did not rescue CFTR, which we expected because it is the increase in Hsp70 expression that promotes maturation of band B to C (19). The inhibition of proteasome-mediated degradation by bortezomib partially rescued ⌬F508-CFTR from ERAD (lane 10). To investigate the effects of bortezomib on components of ERAD, IB3-1 cells were treated with bortezomib for 6 h and then immunoblotted for ERAD components. Induction of Hsp70 and inhibition of VCP were observed with no change in gp78 and Hsp40 (Fig. 3A).
p97/VCP Interacts with the CFTR Immunocomplex-To confirm that both ⌬F508-CFTR and wt-CFTR are VCP substrates, we immunoprecipitated CFTR from IB3-1 and S9 total protein extracts (500 g/ml) using rabbit anti-CFTR polyclonal antibody 169 (15). VCP was co-immunoprecipitated with both ⌬F508-CFTR and wt-CFTR (Fig. 3B, upper  panel), indicating VCP to be the integral component of ERAD of CFTR. It is possible that VCP interacts with residual ⌬F508-CFTR in the S9 cell line. Anti-VCP antibody was used as a positive control, and preimmune serum was used as a negative control for immunoprecipitation from IB3-1 cells. We also observed that VCP was co-immunoprecipitated with Hsp40 probably as part of the ⌬F508-CFTR immunocomplex from IB3-1 cells (Fig. 3C).
To understand the mechanism of VCP-mediated ERAD of CFTR, we analyzed the interaction of VCP with other molecular chaperones known to be associated with ⌬F508-CFTR or wt-CFTR degradation machinery. We immunoprecipitated CFTR, Hsp40, Hsp90, Hsc70, Hsp70, CHIP, and VCP from IB3-1 and S9 total protein extracts (500 g/ml) to evaluate relative VCP levels from each IB3-1 and S9 immunoprecipitate. We found that VCP was independently co-immunoprecipitated with ⌬F508-CFTR, Hsp40, Hsp90, Hsc70, Hsp70, and CHIP from IB3-1 cells (Fig. 4A). Relatively higher amounts of VCP were pulled down with anti-Hsp40 and anti-Hsc70 antisera compared with other proteins, suggesting a stronger interaction or higher stoichiometry. VCP was pulled down together with Hsp90 and Hsp70 immunocomplexes from IB3-1 cells, but not from S9 cells (Fig. 4A). Moreover, VCP had a higher stoichiometry for these molecular chaperones in the presence of ⌬F508-CFTR (IB3-1 cells) compared with wt-CFTR (S9 cells). Freshly isolated bronchial epithelial cells brushed from CF patients homozygous for ⌬F508-CFTR were examined by immunoprecipitation of individual components of ERAD and immunoblotting for VCP or Hsc70 (Fig. 4B). Again, VCP and Hsc70 were prominent with anti-Hsp40 antibody.
VCP Inhibition and Bortezomib Treatment Induce Chloride Efflux-To confirm that ⌬F508-CFTR rescue from ERAD leads to functional cellsurface chloride channels, we measured cAMP-stimulated halide efflux in IB3-1 cells. The IB3-1 cells were transfected with ⌬F508-CFTR-GFP and cotransfected with ⌬VCP, VCPQQ, ⌬gp78, or gp78⌬C or treated with bortezomib (1, 5, or 10 M) for 6 h. We observed minimal or no change in MQAE fluorescence (indicative of Cl Ϫ channel activity) when stimulated by forskolin, although there was a significant change in fluorescence upon genistein treatment (Fig. 5A). Genistein is an isoflavonoid that directly activates CFTR chloride channels (20). We confirmed that the lack of forskolin-mediated CFTR channel activity in IB3-1 cells was due to the presence of endogenous cAMP by using CFTR inhibitor 172 (12) or the cAMP-specific inhibitor myristoylated protein kinase A inhibitor 14 -22 amide (data not shown). We observed that acute exposure to genistein augmented chloride efflux with ⌬VCP or ⌬gp78 and upon treatment with 5 or 10 M bortezomib, suggesting a surface localization for ⌬F508-CFTR (Fig. 5A). These results indicate that inhibiting VCP/gp78 or proteasome activity by bortezomib can rescue functional ⌬F508-CFTR to the cell surface, resulting in a functional increase in cell-surface chloride transport detectable with agonist. Moreover, bortezomib/⌬VCP and genistein had a synergistic effect on the rescue of functional ⌬F508-CFTR from ERAD (Fig. 5A), probably by improving both CFTR trafficking and function. VCP inhibition rescued functional ⌬F508-CFTR by t ϭ 8 min (Fig. 5B). Together, these data demonstrate that we were able to express functional ⌬F508-CFTR in IB3-1 cells by bortezomib treatment or selective inhibition of VCP.
To confirm that down-regulation of IL8 levels is NFB-mediated, we evaluated NFB binding and IL8 levels in IB3-1 cells. IB3-1 cells were transfected (24 h) with ⌬VCP, ⌬F508-CFTR, or ⌬CHOP or treated with 5 M bortezomib (6 h), 5 M MLN-273 (6 h; a novel small molecule proteasome inhibitor), or 5 mM 4-PBA (24 h). CHOP shRNA-mediated CHOP inhibition (⌬CHOP) was used to suppress the oxidative stress in response to VCP or proteasome inhibition. Cells were induced with IL1␤ (1 ng/ml) for 12 h. The supernatants and nuclear extracts from the same experiment were used to determine IL8 levels and NFB binding. NFB binding and IL8 levels were significantly inhibited by ⌬VCP, MLN-273, and bortezomib compared with the control, ⌬CHOP, and 4-PBA. ⌬CHOP and bortezomib had a synergistic effect on NFB binding and IL8 down-regulation (Fig. 6, B and C). Our results indicate that VCP or proteasome inhibition seems to be the most promising therapeutic strategy to restore CFTR trafficking and to inhibit IB␣ degradation simultaneously (16,17) and hence NFB-mediated IL8 activation.

DISCUSSION
The ubiquitin-proteasome system consists of two steps. First, the target protein is conjugated with multi-ubiquitin molecules, which mark the substrate for destruction; second, the target protein is transferred to the 26 S proteasome, unfolded, and degraded. VCP is a multiubiquitin chain-targeting factor that is required for degradation of numerous ubiquitin-proteasome system substrates (24,25). VCP is known to interact with the U-box domain-containing protein Ufd2 and gp78 (ERAD-related ubiquitin-protein isopeptide ligase), resulting in ubiquitination of misfolded protein, followed by retranslocation to the proteasome (8,26,27). In our experiments, inhibition of VCP led to accumulation of CFTR in the ER, and a portion of the rescued mutant CFTR attained the C conformation, indicating passage through the Golgi. Parallel studies from other groups using yeast and mammalian systems demonstrated that Cdc48/VCP functions with the Ufd1-Nlp4 co-chaperone complex to bind ubiquitinated ERAD substrates and to deliver them to proteasomes (9, 28 -30). The Cdc48/VCP chaperone complex binds ubiquitinated misfolded proteins and assists in their delivery from Hsp70 and Hsp90 to the proteasome (28,29). The N-ethylmaleimide-sensitive factor and VCP are related AAA ATPase proteins implicated in membrane trafficking and organelle biogenesis. Although it was recently reported that VCP does not affect membrane trafficking (30), our observation of partial CFTR rescue by VCP inhibition indicates the following possibilities: 1) a small proportion of endogenous VCP can support the membrane trafficking function of specific proteins, but not ERAD; 2) the N-ethylmaleimide-sensitive factor or other related membrane trafficking components become active in response to accumulation of protein(s) in the ER; or 3) VCP inhibition may rescue trafficking proteins from proteasome-mediated degradation.
Two representative mammalian U-box proteins, Ufd2 and CHIP, interact with the molecular chaperones VCP and either Hsp90 or Hsc70, respectively, and are implicated in the degradation of damaged proteins (31). The combination of CHIP with Hsp90 mediates ubiquitylation of the glucocorticoid receptor, and CHIP together with Hsc70 is known to target immature CFTR for proteasome degradation (32,33). We have shown here that VCP is directly associ- ated with CFTR and requisite for ERAD of CFTR, but it is not clear if both VCP-and CHIP-mediated pathways operate parallel to each other. Certainly VCP is also involved in retrotranslocation of tagged protein from the ER to the proteasome, indicating that it may be involved in steps after CHIP.
Inhibition of proteasome activity by N-acetyl-leucyl-leucyl-norleucinal (ALLN, 10 mg/ml) blocks ⌬F508-CFTR degradation and drives accumulation of ubiquitinated forms of ⌬F508-CFTR in Triton X-100insoluble aggregates (34). It was reported recently that inhibition of the proteasome leads polyubiquitinated ⌬F508-CFTR to aggregate because it can be extracted from the ER membrane by the VCP-Ufd1-Npl4 complex (7), but cannot be degraded by the proteasome (34). On the other hand, non-ubiquitinated ⌬F508-CFTR that accumulates in response to inhibition of the Hsp70/CHIP ubiquitin-protein isopeptide ligase does not aggregate because it is inserted into the ER membrane and is bound by cytosolic Hsc70. Thus, whereas ⌬F508-CFTR has a folding defect that prevents it from passing quality control and escaping the ER, it does not appear to be overly aggregation-prone, and cellular chaperones can maintain it in a foldable state (35). The nature of the folding defect of ⌬F508-CFTR and what causes it to be selected for proteasome degradation are not completely clear. To block the VCP-mediated proteasome degradation of CFTR, we used bortezomib. Proteasome inhibition with bortezomib or VCP inhibition partially rescued mature ⌬F508-CFTR in IB3-1 cells. Bortezomib induced Hsp70 and down-regulated VCP protein expression. A recent report also demonstrated the inhibition of VCP by another proteasome inhibitor, hemin (36). Proteasome inhibition may also rescue trafficking protein(s) from degradation that may be involved in ⌬F508-CFTR rescue to the cell surface.
The recent approval of bortezomib as a drug for refractory multiple myeloma (37)(38)(39) marks the beginning of a new era of regulation of protein catabolism for therapeutics. Bortezomib (pyrazylcarbonyl-Phe-Leu-boronate) is an extremely potent, stable, reversible, and selective inhibitor of chymotryptic threonine protease activity (37). Bortezomib shows encouraging results when employed in hematological cancers and solid tumors. Bortezomib selectively induces NFB-mediated apoptosis in cancer cells, whereas normal cells recover from proteasome inhibition (40). Proteasome modulators have been shown recently to be of dual therapeutic importance in pharmacogene therapy of cystic fibrosis airways. The proteasome-modulating agents N-acetyl-leucyl-leucylnorleucinal and doxorubicin enhance CFTR gene delivery and hence CFTR-mediated short circuit current. Moreover, these proteasome modulators also inhibit functional epithelial Na ϩ channel activity and currents independent of CFTR vector administration (41). We have shown here that bortezomib can rescue ⌬F508-CFTR from ERAD and results in the appearance of mature CFTR. A main concern in considering the proteasome as a therapeutic target is the theoretical risk that multiple processes may be affected by proteasome inhibitors. ERAD is a central element of the secretory pathway and has major implications for the generation of various human diseases. ERAD was originally thought to exclusively degrade inefficiently folded and orphan secretory proteins. It is becoming clear that ERAD has important consequences for diverse aspects of cell physiology, including protein folding and transport, metabolic regulation, immune response, and ubiquitinproteasome-dependent degradation (2). The risk of global suppression of proteasome degradation may be balanced by the favorable anti-inflammatory effect we observed.
Several studies have indicated that pulmonary inflammation may occur early in the course of CF, although the existence of primary inflammation is controversial (42,43). It has been shown that human bronchial epithelial cells expressing ⌬F508-CFTR generate more IL8 in response to tumor necrosis factor-␣, IL1␤, or P. aeruginosa (21). These studies indicate that neutrophilic association with increased IL8 in the airways is a prominent early feature of CF and suggest that the exuberant airway inflammation is a component of the CF phenotype. In quiescent cells, NFB is sequestered to the cytoplasm by its interaction with a member of the IB inhibitory family that includes IB␣ and IB␤. After cell stimulation, IB␣ is phosphorylated, polyubiquitinated, and degraded by 26 S proteasome activity. IB␣ degradation unmasks nuclear localization signals that allow NFB to be transported to the nucleus, resulting in activation of IL8 gene transcription (23). VCP inhibition and 5 m bortezomib significantly inhibited IL1␤-induced IL8 levels in IB3-1 cells from 100.49 Ϯ 0.95 pg/ml (control) to 89.96 Ϯ 1.43 (⌬VCP) and 87.27 Ϯ 2.05 pg/ml (bortezomib) (p Ͻ 0.01). VCP physically associates with IB␣ and the 26 S proteasome and targets IB␣ to the proteasome for degradation (16). The IL8 inhibition by VCP interference can be explained by IB␣-mediated NFB inhibition. Bortezomib can enter mammalian cells and inhibit NFB activation and NFB-dependent gene expression. Bortezomib also inhibits tumor necrosis factor-␣-induced gene expression of the cell-surface adhesion molecules E-selectin, ICAM-1 (intercellular adhesion molecule 1), and VCAM-1 (vascular cell adhesion molecule 1) on primary human umbilical vein endothelial cells (44,45). In a rat model of streptococcal cell wall-induced polyarthritis (46), bortezomib attenuates the neutrophil-predominant acute phase and markedly inhibits the progression of the T cell-dependent chronic phase of the inflammatory response (44).
Proteins unable to fold correctly cause ER stress and activate the UPR. The proteasome contributes to ERAD by relieving the ER stress. Accumulation of unfolded protein within the lumen of the ER leads to prolonged UPR activation, which in turn causes oxidative stress and finally cell death (47). To target the oxidative stress, we inhibited the major stress-inducible transcription factor CHOP/GADD153, which serves as a pro-apoptotic signal in response to the UPR (Fig. 7). The proteasome inhibitor is known to induce CHOP (48), although apoptosis was not apparent after proteasome or VCP inhibition in the present study. VCP itself was verified to be the major component of ER stress-induced apoptosis (47); hence, lack of apoptosis by VCP inhibition is apparent. Proteasome inhibition blocks NFB activation in response to exposure to cytokines or UV irradiation by stabilizing the labile IB␣ inhibitory protein (49). CHOP inhibition together with proteasome inhibition is synergistic in repression of NFB-mediated IL8 activation. Central to the pro-apoptotic function of proteasome inhibition is the transcriptional regulator CHOP, whereas VCP itself mediate ER stress-induced apoptosis.
VCP inhibition was more efficient in the rescue of the C form of CFTR compared with gp78 or proteasome inhibition. Mutant CFTR was relatively inert to stimulation by cAMP-mediated phosphorylation, but very sensitive to genistein, which directly activates both mutant and wild-type CFTRs. Because VCP is required for retrotranslocation of CFTR from the ER, VCP inhibition retains CFTR in the ER for trafficking to the plasma membrane. The gp78-mediated ubiquitination and proteasome degradation of CFTR occur after retrotranslocation from the ER; therefore, inhibition of these events is less effective in retaining CFTR in the ER. We conclude that VCP is the integral component of ERAD and ER stress pathways induced by the UPR in CF and may be central to the efficacy of CF drugs that target the ubiquitin-proteasome system. Modulating proteasome degradation by bortezomib or VCP shRNA rescues functional mutant CFTR to the cell surface and suppresses NFB-mediated IL8 activation. This ability to ameliorate secondary aspects of CF disease pathophysiology in addition to rescue of CFTR to the cell surface is promising for CF therapeutics. Moreover, the identification and selective modulation of ERAD components as potential therapeutics for a wide range of human diseases associated with ERAD promise an exciting and innovative area of investigation.  . Inhibition of the ubiquitin-proteasome pathway in CF. Proteasome or VCP inhibition stabilizes IB and blocks NFB-mediated IL8 activation in addition to rescuing CFTR from degradation. Inhibition of the ubiquitin-proteasome pathway reduces general translation to avoid overload of the protein folding machinery, although it enhances translation and transcription of specific transcription factors such as heat shock proteins (HSPs) to enhance protein folding and assembly and CHOP to induce oxidative stress and apoptosis.