Several mutations within the BRICHOS domain of surfactant protein C (SP-C) have been linked to interstitial lung disease. Recent studies have suggested that these mutations cause misfolding of the proprotein (proSP-C), which initiates the unfolded protein response to resolve improper folding or promote protein degradation. We have reported that in vitro expression of one of these proteins, the exon 4 deletion mutant (hSP-CΔexon4), causes endoplasmic reticulum (ER) stress, inhibits proteasome function, and activates caspase-3-mediated apoptosis. To further elucidate mechanisms and common pathways for cellular dysfunction, various assays were performed by transiently expressing two SP-C BRICHOS domain mutant (BRISPC) proteins (hSP-CΔexon4, hSP-CL188Q) and control proteins in lung epithelium-derived A549 and kidney epithelium-derived (HEK-293) GFPu-1 cell lines. Compared with controls, cells expressing either BRICHOS mutant protein consistently exhibited increased formation of insoluble aggregates, enhanced promotion of inositol-requiring enzyme 1-dependent splicing of X-box binding protein-1 (XBP-1), significant inhibition of proteasome activity, enhanced induction of mitochondrial cytochrome c release, and increased activations of caspase-4 and caspase-3, leading to apoptosis. These results suggest common cellular responses, including initiation of cell-death signaling pathways, to these lung disease-associated BRISPC proteins.
- surfactant protein C
- misfolding proteins
- protein aggregation
- endoplasmic reticulum stress
- proteasome inhibition
surfactant protein c (SP-C) is a unique component of pulmonary surfactant, a heterogeneous complex of lipids and proteins that lowers lung surface tension as well as contributes to the pulmonary host defense system. The presence of SP-C in surfactant enhances adsorption and spreading of surfactant phospholipids, thereby promoting surface film formation. Mature SP-C is a hydrophobic peptide with an Mr of 3,700 produced from a 21-kDa bitopic precursor transmembrane protein (proSP-C) that is synthesized and proteolytically processed exclusively in alveolar type II (ATII) cells (1). During subcellular trafficking, proSP-C undergoes homomeric association mediated by the mature domain (51). SP-C is released into the alveolar space via regulated secretion together with phospholipids and other proteins contained in lamellar bodies (specialized secretory organelles of ATII cells).
Multiple reports have described an association between mutations in the gene encoding SP-C (SFTPC) and chronic interstitial lung diseases (ILDs) in both infants and adults (3, 37, 48, 49). Several of these mutations result in amino acid substitutions clustered within the proprotein's COOH-terminal domain, known as BRICHOS, a region of ∼100 residues with structural homology to a number of proteins associated with degenerative and proliferative diseases in other organs (46). A deletion mutation (hSP-CΔexon4) has also been described in the intronic region of SFTPC immediately upstream of exon 4 that results in alternative splicing of the proSP-C mRNA and exclusion of exon 4 (38). This event produces a defective precursor protein in which the COOH-terminal domain of proSP-C is foreshortened by 37 amino acids and removes the putative disulfide bond-forming cysteine residue that is believed to be responsible for folding and consequent processing of proSP-C (21). Consistent with the characteristics of other misfolded proteins, unprocessed mutant peptide accumulation in ubiquitinated perinuclear inclusion bodies with subsequent aggresome formation is observed in various lung epithelial cell lines transiently transfected with hSP-CΔexon4 (33, 50). Ubiquitin/proteasome activity is also inhibited in cells expressing hSP-CΔexon4 (33). Furthermore, expression of the mutant protein is retained in the endoplasmic reticulum (ER) and can elicit the unfolded protein response with concomitant production of multiple ER stress species and induction of apoptotic cell death (4, 5, 33). In addition, the presence of proSP-C in ATII cells and the absence of detectable mature SP-C in the alveolar space of patients with heterozygous expression of the hSP-CΔexon4 mutation suggest a dominant negative effect.
A SFTPC mutation that substitutes a conserved leucine residue for glutamine within the BRICHOS domain (hSP-CL188Q) has also been identified in association with familial ILD with pathological features comparable to patients with the hSP-CΔexon4 isoform (49) (Fig. 1A). Immunohistochemical and electron micrograph evaluations of patients’ lung tissue show atypical ATII cells with abnormal lamellar bodies and aberrant subcellular localization of the precursor protein. Transient expression of the mutant proprotein in mouse lung epithelial cell lines has also been shown to induce cell toxicity (49). The structural configuration (1, 18, 21, 51, 53) of proSP-C, coupled with the location of this missense mutation adjacent to a conserved cysteine residue at 189, suggests that the hSP-CL188Q mutation may affect proSP-C folding. This could be achieved via hindrance of disulfide bond formation by the adjacent cysteine 189 residue and/or destabilizing the conformation of the propeptide (17). This is graphically illustrated in Fig. 1A. Moreover, similarities in the clinical manifestations of hSP-CΔexon4 and hSP-CL188Q mutations, together with experimental data available thus far, support the notion that these two mutations could share common pathways for cellular injury and death. We therefore hypothesized that expression of hSP-CL188Q causes apoptotic cell death via multiple cellular responses. We report presently that the two SP-C BRICHOS domain mutant (BRISPC) proteins not only elicit similar cellular responses including induction of ER stress, disruption of the ubiquitin proteasome pathway, deposition of toxic cellular aggregates, and cell death, but also induce previously undescribed (with BRICHOS domain mutations) death-promoting pathways for apoptosis involving caspase-4 and mitochondrial signaling.
MATERIALS AND METHODS
SP-C cDNA expression.
Human proSP-CWT (amino acids 1–197) in pEGFPC1 and pcDNA3 vectors has been previously described, as has proSP-CΔexon4 (50, 51). To create EGFP·hSP-CL188Q, proSP-CWT in pcDNA3 vector was used as a template in a single PCR reaction with two primers, dTCCGGACTCAGATCTATGGATGTGGGCAGCAAAGAGGTCCTG (forward) and dGGAGGCGTCGGATCCCTAGATGTAGTAGAGCGGCACCTCGCCACACTGGGTGCTCACGGCCATGCCCAG (reverse). These primers respectively contain BglII and BamHI restriction enzyme recognition sites for insertion into the pEGFPC1 vector. Human proSP-CWT, proSP-CΔexon4, and proSP-CL188Q in DsRed vector were created using the enhanced green fluorescent protein (EGFP)-tagged wild-type and mutant proSP-C plasmids as templates and a single PCR reaction with two primers. The primers used were dGGACTCAGATCTCGAGAAATGGATGTGGGCAGCAAAGAG (forward) and dTGTGGTATGGGTACCTATGATCAGTTATCTAGATCC (reverse) containing XhoI and KpnI restriction enzyme recognition sites, respectively, for insertion into the DsRed vector. Sequence fidelity of each PCR product was verified by automated DNA sequencing (in both directions) performed at the Core Facility in the Department of Genetics at the University of Pennsylvania. Primers were designed to amplify ∼600 bases.
Human lung epithelium (A549)- and kidney epithelium (HEK-293)-derived GFPu-1 cell lines were obtained from American Type Culture Collection (Manassas, VA).
Immunocytochemistry and fluorescence imaging was performed according to the previously described protocol (33). Following permeabilization, cells on coverslips were immunolabeled with primary antibodies for 1 h at room temperature at the following dilutions: anti-CD63 (Immunotech, Marseilles, France), 1:1,000; anti-calnexin (Stressgen, Victoria, Canada), 1:200; anti-EEA1 (Affinity BioReagents, Golden, CO), 1:100; and anti-ubiquitin (FK2; Affiniti, Devon, UK), 1:100. Texas red-conjugated secondary goat anti-mouse IgG monoclonal or secondary goat anti-rabbit IgG polyclonal antibodies (Jackson ImmunoResearch Laboratories, West Grove, PA) were used for visualization at 1:200 dilution.
In situ apoptotic cell death assays.
Two separate assays were used for determining apoptosis: 1) activated caspase-3 labeling in permeabilized cells (see Immunocytochemistry) and 2) annexin V binding. The assay for annexin V binding was performed according to the manufacturer's recommendation (Roche Diagnostics, Penzberg, Germany) with minor modifications. A549 cells grown to 70% confluence in 35-mm glass-bottom dishes [with 0.16- to 0.19-mm cover glasses (MatTek, Ashland, MA) for live microscopic observation] were transiently transfected with EGFP/proSP-C constructs (4 μg/dish) using FuGene 6. For positive control, cells in separate 35-mm glass-bottom dishes were treated with TNF-α (BD Biosciences, Bedford, MA) overnight. Forty-eight hours following the introduction of plasmid DNA, cells were washed twice (5 min each) with phosphate-buffered saline (PBS: 137 mM NaCl, 10 mM Na2HPO4, 2.7 mM KCl, 1.8 mM KH2PO4, pH 7.2), incubated with incubation buffer (10 mM HEPES/NaOH, pH 7.2, 140 mM NaCl, 5 mM CaCl2) for 10 min, and incubated with annexin V-Alexa647 labeling solution [incubation buffer plus 10 μl of annexin V-Alexa647 (Roche Diagnostics)] for 15 min at room temperature. To control for necrosis, cells were monitored for propidium iodide nuclear labeling by adding 10 μg/ml propidium iodide concomitantly with the annexin V. Cells were visualized with a confocal fluorescence microscope (Olympus, Center Valley, PA).
Immunoblotting was performed according to the previously described procedure (33) using successive incubations with either primary polyclonal antisera [anti-GFP (Molecular Probes, Eugene, OR), 1:5,000; anti-FLAG (Rockland Immunochemicals, Gilbertsville, PA), 1:1,000; anti-activated caspase-3 (Sigma, St. Louis, MO), 1:200; anti-cytochrome c (BD Bioscience, San Jose, CA), 1:100; anti-β-actin (Sigma), 1:2,000] or primary monoclonal antibodies [MTCO2 (mitochondrial marker; AbCam, Cambridge, MA), 1:200; anti-caspase-4 (Stressgen, Victoria, Canada), 1:500] followed by either goat anti-rabbit or goat anti-mouse horseradish peroxidase-conjugated secondary antibody, respectively (1:10,000). Bands were visualized by enhanced chemiluminescence using a commercially available kit (ECL Western blotting detection reagents; Amersham, Arlington Heights, IL). Chemiluminescent images were produced either by exposure to film or by direct acquisition using the Kodak 440 imaging system (New Haven, CT).
In vitro proteasome inhibition assay and induction of ER stress and apoptosis.
A proteasome inhibition assay has been described previously (33). To induce ER stress and/or apoptosis, cells were treated overnight, at the indicated concentrations, with either tunicamycin Streptomyces sp. or thapsigargin (both from Sigma) for ER stress, TNF-α (BD Biosciences) for caspase-3 activation and apoptosis, or etoposide (Sigma) for ER stress-independent induction of apoptosis.
Mitochondrial cytochrome c release.
Determination of mitochondrial cytochrome c release to the cytosol was performed according to a previously described procedure with minor modification (40). Cell pellets collected by scraping dishes and centrifuging at 300 g were washed twice with 500 μl of PBS. For protein enrichment, cells from two separate 30-mm culture dishes were used per fraction. Pellets were resuspended in 150 μl of ice-cold permeabilizing buffer (75 mM KCl, 1 mM NaH2HPO4, 250 mM sucrose, 230 μg/ml digitonin) containing protease inhibitors and incubated on ice for 15 min for plasma membrane permeabilization and release of soluble cytosolic components. To monitor the level of cell permeability, a trypan blue dye exclusion assay was used on sample aliquots. Cells were centrifuged at 5,000 g, 4°C for 5 min. Supernatant was transferred into a new tube and centrifuged again. The supernatant of this last centrifugation contained the cytosolic fraction, which was mixed with an equal volume of 2× RIPA buffer (1× RIPA buffer is 1% Igepal, 0.5% sodium deoxycholate, 0.1% SDS, 0.2 mM sodium orthovanadate, and 50 mM sodium fluoride, in PBS) containing protease inhibitors. For the organelle fraction including the mitochondria, the pellet from the first centrifugation following permeabilization was washed twice with 500 μl of PBS (6,000 g, 4°C) for 5 min and resuspended in 40 μl of 1× RIPA buffer.
Triton X-soluble and -insoluble fractionation assay.
Fractionation assay to separate Triton X-soluble from Triton X-insoluble proteins was performed as previously described (31) with some modification. A549 cells in 30-mm culture dishes were transiently transfected with EGFP-tagged proSP-C constructs (10 μg/dish) using the calcium phosphate method. At 72 h posttransfection, cells were washed twice with 500 μl of ice-cold PBS and lysed in ice-cold lysis buffer [50 mM Tris·HCl (pH 7.2), 150 mM NaCl, 2 mM EDTA, 1% (vol/vol) Triton X-100, and .05% (vol/vol) sodium deoxycholate] containing protease inhibitors. For protein enrichment, cells from two separate 30-mm culture dishes were used per fraction. Lysed cells were sedimented at 16,000 g for 30 min at 4°C. The supernatants were transferred into new tubes and centrifuged again at 16,000 g for 30 min at 4°C. The supernatants from the second centrifugation were saved as the Triton X-soluble fraction. Pellets from the initial fractionation were washed twice with lysis buffer and solubilized with SDS buffer (lysis buffer containing 1% SDS in place of Triton X-100) and saved as the Triton X-insoluble fraction. SDS-PAGE sample buffer was added to both the supernatant and pellet fractions for subsequent electrophoresis.
Experimental data were analyzed by one-way analysis of variance with the Tukey-Kramer post hoc test using GraphPad InStat software, version 3.0 for Windows (GraphPad Software, San Diego, CA). All values are means ± SE.
The SP-CL188Q mutant is trafficked to ubiquitinated cytosolic inclusion bodies and perinuclear aggregate.
We have reported previously that the hSP-CΔexon4 BRISPC propeptide localizes to both the ER and in ubiquitinated perinuclear aggregates when transiently expressed in various epithelial cell lines including A549, MLE15, and HEK-293 cells (33). Moreover, we have shown that non-BRICHOS domain mutations of SP-C, including SP-CE66K and SP-CI73T, are targeted to early endosome compartments (3, 48). To determine the trafficking pattern of hSP-CL188Q, we transfected A549 cells with EGFP-tagged wild-type (EGFP·hSP-CWT) and mutant (EGFP·hSP-CL188Q) isoforms of SP-C. At 48 h posttransfection, EGFP·hSP-CWT was localized to CD63 (a marker for lamellar bodies and lysosomes)-positive and EEA1 (an early endosome marker)-negative vesicles (Fig. 1B), the expected target vesicle for the wild-type isoform of SP-C. In contrast, EGFP·hSP-CL188Q was rarely observed in CD63-positive vesicles (Fig. 1C, left, arrows) and was not detected in EEA1-containing vesicles (Fig. 1C, right). Instead, three distinct patterns of trafficking are apparent predominantly localizing EGFP·hSP-CL188Q to the calnexin-positive ER compartment (Fig. 1D, left) and ubiquitin-positive inclusions with punctate cytosolic distribution and perinuclear-confined localization (Fig. 1D, center and right). Quantitative analyses of these hSP-C isoforms revealed that EGFP·hSP-CWT had minimal colocalization with either ubiquitin or calnexin (Fig. 1E). By contrast, both EGFP·hSP-CΔexon4 and EGFP·hSP-CL188Q showed a significant amount of colocalization with calnexin (>20%) and ubiquitin (>40%) (Fig. 1E).
BRISPCs form insoluble cytosolic aggregates.
To date, direct biochemical evidence of solubility of BRISPC proteins has not been described. To determine the susceptibility of these proteins to form insoluble aggregates, lysates from A549 cells transfected with various isoforms of EGFP-tagged SP-C proteins or vector alone (EGFPC1) were partitioned into detergent-soluble and -insoluble fractions (Fig. 1F). Immunoblots of SDS-PAGE-separated proteins from cell lysates harvested 72 h following transfection showed that EGFP·hSP-CWT and EGFPC1 were almost exclusively partitioned in soluble fractions. In contrast, a substantial portion of either EGFP·hSP-CΔexon4 or EGFP·hSP-CL188Q proteins segregated with Triton X-insoluble fractions (Fig. 1F), indicating a tendency for these BRISPC proteins to form insoluble aggregates.
Expression of hSP-CL188Q inhibits ubiquitin/proteasome function.
Accumulation of protein aggregates suggests an ubiquitin/proteasome system (UPS) that is overwhelmed by chronic production of misfolded proteins and is unable to sustain their degradation (20, 26). To address whether expression of hSP-CL188Q inhibits the UPS, we used GFPu-1 cells. GFPu-1 is a human kidney epithelium (HEK-293)-derived, stably transfected cell line expressing a short degron (CL1) tagged to GFP (10). GFP stability in these cells can be used as a marker to determine relationships between protein expression and UPS dysfunction (2). Under normal conditions, the GFP/degron expression is short-lived, since the efficient UPS of HEK cells rapidly degrades the product soon after translation, before any detectable green fluorescence is observed (Fig. 2A, bottom left). However, conditions that impair the UPS, thereby inhibiting proteasome degradation, will promote the accumulation of the GFP-tagged peptide. Control treatment with lactacystin, a proteasome inhibitor, resulted in GFP expression across the entire cell culture population (Fig. 2A, bottom center). Cells transiently transfected with DsRed-tagged hSP-CWT (Fig. 2A, top left) showed no GFP fluorescence (Fig. 2A, top center), indicating that the UPS is not affected by the presence of hSP-CWT. In contrast, cells expressing DsRed·hSP-CL188Q (Fig. 2A, middle left) frequently exhibited GFP fluorescence (Fig. 2A, middle center). Immunoblot of GFPu-1 whole cell lysates using anti-GFP showed a dose-dependent increase in GFP protein levels in samples where cells were treated with lactacystin (Fig. 2B). In addition, significant increases in GFP protein levels were observed in GFPu-1 whole cell lysates transfected with either hSP-CΔexon4 or hSP-CL188Q compared with those transfected with vector or hSP-CWT (Fig. 2B). Together, these results demonstrate the capacity of BRISPC protein expression to inhibit the UPS.
Aggregate prone BRISPCs induce apoptotic cell death via activation of caspase-3.
It has been shown in other aggregate prone proteins that mutant protein expression results in apoptosis via activation of caspase-3 (27, 33). To determine whether expression of hSP-CL188Q elicits a similar response, we assayed A549 cells transfected with EGFP·hSP-CL188Q for apoptosis and caspase-3 activation by fluorescence labeling. Plasma membranes of live cells containing mutant aggregates were frequently labeled with annexin V (Fig. 3A, bottom), an agent that binds translocated (inner to outer leaflet) plasma membrane phosphatidylserine during apoptosis. The absence of nuclear labeling by propidium iodide in these cells (Fig. 3A, bottom) substantiates apoptotic and not necrotic cell death. Caspase-3 labeling (using an antibody that recognizes the activated form of caspase-3) was also more frequently observed in cells containing EGFP·hSP-CL188Q (Fig. 3, A, top, and B) than in those transfected with EGFP·hSP-CWT (Fig. 3B).
Caspase-3 activation by BRISPC proteins was confirmed using whole cell lysate immunoblot analysis (Fig. 3C), where cell lysates from A549 cells treated with positive control TNF-α, a known inducer of apoptosis in A549 cells, showed caspase-3 activation in a dose-dependent manner. Moreover, 48 h following transfection, significant dose-dependent activation of caspase-3 was observed in both hSP-CΔexon4- and hSP-CL188Q-expressing cells compared with cells expressing either hSP-CWT or vector (Fig. 3C).
BRISPC expression induces inositol-requiring enzyme 1-dependent XBP-1 mRNA splicing.
Accumulation of unfolded proteins in the ER cells activates an intracellular signaling pathway termed the unfolded protein response (UPR). One of the major proximal sensors of the UPR is inositol-requiring enzyme 1 (IRE1), an ER transmembrane protein kinase/endoribonuclease that initiates the UPR signal by regulating synthesis of transcription factors such as the X-box binding protein-1 (XBP-1) through XBP-1 mRNA splicing. XBP-1 in turn regulates ER-resident chaperone genes such as BiP/GRP78, HEDJ, EDEM, and HDJ-2/HSP40 (28).
To determine ER involvement in response to BRISPC protein expression, we utilized the ERAI plasmid construct, which has been used as an indicator for ER stress (14). The ERAI construct contains a partial sequence of XBP-1, including the 26-nucleotide (nt) ER stress-specific intron, fused to the gene encoding venus (a variant of green fluorescent protein) with a FLAG tag at the NH2 terminus. Under normal conditions, a stop codon that is added between the XBP-1 and venus sequences eliminates the venus florescence expression. However, during ER stress, the 26-nt intron is spliced out by IRE1, resulting in the frame shift of the fusion mRNA and the expression of the fluorescent protein.
IRE1 participation was assessed using a strategy in which various isoforms of DsRed-tagged proSP-C plasmid constructs were cotransfected with the ERAI construct in A549 cells. This particular ERAI plasmid lacks the DNA-binding domain (DBD) of XBP-1 (XBP-1ΔDBD), which promotes nuclear translocation. Consequently, expression of the fusion proteins remains cytosolic. Compared with DsRed·hSP-CWT, cells transfected with DsRed·hSP-CL188Q consistently expressed the XBP-1·venus fusion protein (Fig. 4A). Similar results were obtained with DsRed·hSP-CΔexon4 (data not shown). Immunoblot analyses of FLAG-tagged XBP-1 expression in cell lysates cotransfected with ERA1 and various isoforms of SP-C confirmed that spliced XBP-1 expression was significantly higher when either BRISPC mutant (hSP-CΔexon4 or hSP-CL188Q) was expressed compared with hSP-CWT or vector alone (Fig. 4B). Moreover, control experiments showed that the ER stress-inducing reagent thapsigargin promoted the expression of spliced XBP-1 in a dose-dependent manner, whereas etoposide, a non-ER stress inducer of apoptosis, had no effect (Fig. 4B). These results suggest that misfolded BRISPC proteins induce ER stress, leading to IRE1-dependent activation of XBP-1.
Caspase-4 is activated by the expression of BRISPCs.
Prolonged ER stress can induce apoptosis via caspase-4 activation (13). Caspase-4 is a member of the caspase-1 subfamily that includes the mouse homolog caspase-12. Localized to the ER membrane, caspase-4 is cleaved during ER stress, leading to the activation of caspase-3 and apoptosis. Caspase-4 involvement during apoptotic induction by BRISPCs was evaluated in transiently transfected A549 cells. Control studies showed that an apoptosis-inducing cytokine (TNF-α) and known ER stressors tunicamycin and thapsigargin induced procaspase-4 cleavage in a dose-dependent manner (Fig. 5A). In contrast, etoposide, a non-ER stress inducer of apoptosis, at a dose imparting a comparable extent of cell death to the ER stressors, failed to cleave the procaspase (Fig. 5A). These results are consistent with previous reports showing equivalent procaspase-4 cleavage in human neuroblastoma (HK-N-SH) and carcinoma (HeLa) cells (13), implying ER stress-specific intrinsic caspase-4 activation in A549 cells. Similarly, procaspase-4 cleavage was significantly induced 48 h posttransfection in hSP-CΔexon4- and hSP-CL188Q- compared with either hSP-CWT- or vector-expressing cells (Fig. 5B), indicating a potential role for caspase-4 in the apoptotic pathway caused by the expression of these misfolded proteins.
Cytochrome c is released into the cytosol in BRISPC-expressing cells.
Cytochrome c is a soluble protein localized to the intermembrane space and peripherally attached to the surface of the inner mitochondrial membrane. In response to a variety of apoptosis-inducing agents, cytochrome c can be released from mitochondria into the cytosol (25) and participates in the signaling cascade that induces caspase-3 activation and apoptosis (30). To evaluate cytochrome c involvement during BRISPC expression, A549 cells either treated with control reagent overnight or transfected with various isoforms of SP-C for 48 h were harvested and partitioned into cytosolic and organelle fractions. The purity of the fractionated samples was confirmed by immunoblotting using a mitochondrial resident protein-specific antibody (Fig. 6, B and E). Although cytochrome c remained predominantly partitioned within the mitochondria-containing organelle fractions (Fig. 6, A and D, bottom), its dose-dependent enrichment in the cytosol fraction was apparent in control experiments when A549 cells were treated with etoposide (Fig. 6, A, top, and C). Treatment of cells with the ER stressor tunicamycin did not elicit cytochrome c release (data not shown). Moreover, the cytosolic fraction of cells expressing either hSP-CΔexon4 or hSP-CL188Q contained a significantly higher quantity of cytochrome c than those expressing either hSP-CWT or vector (Fig. 6D, top, and F). Together, these results suggest induction of ER stress-independent mitochondrial cytochrome c release by the expression of the BRISPC proteins.
We have previously demonstrated that one BRISPC protein (hSP-CΔexon4) induces the UPR and apoptosis in a manner similar to that described for misfolded proteins in other systems (20, 22, 26, 27). The present work extends these observations by linking UPR to apoptosis, identifying an additional apoptotic pathway mediated by mitochondrial dysfunction, and characterizing cell death signaling mechanisms common to the expression of two well-known BRISPC proteins that may extend to other BRISPC isoforms.
Inhibition of proteasome function by both BRISPC isoforms was demonstrated using GFPu-1 cells (Fig. 3), suggesting a direct cause-and-effect relationship between BRISPC mutants and disruption of the UPS as proposed for other misfolded mutant proteins associated with conformational diseases. Other aggregate-prone misfolded mutant proteins linked to conformational diseases, such as the polyglutamine repeat expansion of the huntingtin protein associated with Huntington disease, the cystic fibrosis transmembrane conductance regulator mutant (CFTRΔF508) linked to cystic fibrosis, and the α-synuclein protein associated with Parkinson's disease, have all been shown to form cytosolic accumulation of their respective misfolded proteins that involves inhibition of the UPS (2, 7, 36, 42). Using GFPu-1 cells, studies have demonstrated that transient expression of the expanded huntingtin proteins and CFTRΔF508 caused almost total inhibition of the UPS (2). The subsequent formation of cytosolic aggregates of these misfolded proteins implies a further decline of UPS function due to a positive feedback mechanism. Mutant α-synuclein expression has been shown to significantly decrease proteasome activity in GFPu-1 and other cells as well as to cause selective toxicity to catecholaminergic neurons when expressed in primary midbrain cultures (7, 42). Moreover, the polyglutamine repeat-associated spinocerebrocellular atrophy protein (SCA3) of Machado-Joseph disease shows several similarities with the BRISPC proteins in terms of cellular response to their expression. In addition to inhibition of proteasome function, expression of SCA3 results in ER stress, formation of cytosolic aggregates, and apoptosis (36). Together, these findings further support common cellular responses to protein expression of BRISPC and other proteins associated with conformational diseases.
The in vitro inhibition of proteasome function by BRISPC proteins is cell model independent. Both A549 cells (33) and HEK-293 cells (5), transiently and constitutively expressing hSP-CΔexon4, respectively, have altered cellular activity. In HEK-293 cells that constitutively express low levels of hSP-CΔexon4, the toxic effect of the mutant propeptide is ameliorated through a cellular adaptive response involving the NF-κB pathway (5). However, when infected with virus, the proteasome function of these stably transfected cells was further inhibited, leading to the accumulation of the propeptide and increased susceptibility to virus-induced cell death. This suggests that environmental factors such as viral infection play an important role in the development and progress of lung diseases associated with BRISPC mutations.
The ER has developed a tightly regulated quality control mechanism such that even the slightest alteration in a protein's makeup that disturbs folding efficacy can cause the recognition of the nascent protein as misfolded, leading to subsequent accumulation and/or retrotranslocation of the protein to the cytosol (22, 45). In the cytosol, misfolded proteins are primarily destined for proteasome-mediated ER-associated degradation (ERAD). However, persistent production of aberrant proteins may shift the balance between synthesis by the ER to degradation of mutant proteins by the proteasome, resulting in accumulation of the misfolded proteins in the form of cytosolic aggregates. We have found that the BRISPC isoforms examined show both ER retention and formation of cytosolic aggregates (33) with an appreciable accumulation of insoluble forms (Fig. 1D). Moreover, we have previously shown time and expression level dependency of hSP-CΔexon4 in the formation of aggregates (33). Similar results can be seen in hSP-CL188Q-expressing cells (Mulugeta S, unpublished observation). However, unlike hSP-CΔexon4, which appeared to be entirely targeted to ERAD or to aggresomes depending on the level of expression (4, 33), a relatively small population of the hSP-CL188Q propeptides were properly targeted in CD63-positive vesicles (Fig. 1C, arrows) and were processed similarly to the wild-type isoform (Fig. 1F). Although the conformational changes of the hSP-CΔexon4 deletion mutation may be severe, missense mutation of hSP-CL188Q may present a more moderate change, sometimes resulting in relatively preserved propeptide folding efficacy. Therefore, it is likely that some hSP-CL188Q mutant proteins are folded or refolded properly, thereby escaping the degradation pathway and resulting in propeptide processing similar to that of the wild-type isoform. This phenomenon has been well characterized for the CFTR protein, where 70–80% of the newly synthesized wild-type CFTR proteins are normally destined to ERAD, whereas the remaining 20–30% are properly transported to the cell surface (15, 52).
ER stress appears to be a consequence of an insufficient ER adaptive response due to persistent production of misfolded proteins (22, 57). The accumulation of these proteins leads to the initiation of apoptosis, which in BRISPC-expressing cells as well as in various other systems appears to be mediated by caspase-4 (13, 34, 56). Localized in the ER membrane, cleavage of this cysteine protease has been shown when cells are specifically treated with ER stress-inducing agents but not with ER stress-independent apoptotic stimuli. Caspase-12, the mouse homolog to human caspase-4, appears to be the major pathway that induces apoptosis in response to ER stress in mice (22, 56). Caspase-12-deficient mice are resistant to ER stress-induced apoptosis, but cells from these mice can be led to apoptosis by non-ER stress apoptotic agents (35). Caspase-4 involvement has also been implicated in the neurodegenerative disorder of Alzheimer's disease. The extracellular plaques (containing amyloid-β peptide) found in the brains of Alzheimer's patients cause neuronal cytotoxicity (55) mediated by caspase-4 (13) and caspase-12 (34) in human and murine cells, respectively. However, recent studies have demonstrated that caspase-4 may not be required for ER stress-induced apoptosis in some systems (39). Using murine and human cell lines that lack the murine caspase-12 and the human caspase-4, respectively, Obeng and Boise (39) have shown that caspase-12/-4 deficiency did not protect the cells from ER stress-induced apoptosis triggered by tunicamycin, thapsigargin, or brefeldin A.
The release of the apoptotic factor cytochrome c from mitochondria by the expression of BRISPC proteins (Fig. 6) raises the possible involvement of mitochondria with at least three subcellular events that include ER stress, impairment of the UPS, and protein aggregates. The role of ER stress in mitochondrial cytochrome c release has been previously described in some cell systems. In rodent cells, mitochondrial cytochrome c is released in a caspase-12-independent (12) and caspase-8-dependent (16) manner when cells are treated with ER stress-inducing agents. In contrast, the mutant form of α-synuclein has been shown to induce ER stress, proteasome dysfunction, and mitochondrial cytochrome c release accompanied by activation of caspases-12 and -3 in differentiated PC12 cells (47), which actually parallels our findings in BRISPC-expressing A549 cells. We have discovered that when A549 cells are treated with the ER stress-inducing agent tunicamycin, cytochrome c is not released (data not shown), implying that BRISPC induction of cytochrome c is an ER stress-independent signaling pathway. We have also demonstrated that BRISPC expression inhibits proteasome function (Fig. 2) and forms detergent-insoluble cytosolic aggregates of the propeptide (Fig. 1). Furthermore, a recent report indicates that although the BRICHOS domain of wild-type proSP-C functions to prevent aggregation of the propeptide, mutation in this domain appears to impede this function (17). Aggregate formation in other systems is also frequently accompanied by mitochondrial cytochrome c release and apoptosis (6, 43, 54). Moreover, intra- and/or extracellular expression of aggregate prone proteins is associated with mitochondrial cytochrome c release in several conformational diseases such as Alzheimer's disease (24, 44), Huntington disease (23), and Parkinson's disease (47). Similar to β-amyloid aggregates of Alzheimer's disease, SP-C can form β-sheet amyloid fibrils (9, 11, 29) in solution comparable to those hSP-C amyloid fibrils found in lung lavage obtained from patients with pulmonary alveolar proteinosis (11). Collectively, these studies suggest that the cytochrome c release observed in BRISPC-expressing A549 cells is likely to be a product of UPS impairment and aggregate formation rather than a consequence of ER stress.
We have previously shown that proSP-C self-associates (via the mature domain) during its biosynthesis as it traffics from the ER to lamellar bodies through the Golgi, small vesicles, multivesicular bodies, and composite bodies (51). We have also demonstrated that hSP-CΔexon4 is transiently retained in the ER, is not proteolytically processed, and ultimately acts as a dominant negative to direct SP-CWT to aggregates (50). In vivo, patients heterozygous for SP-CΔexon4 do not express mature SP-C (38). Such a phenomenon not only inhibits the wild-type isoform from being processed and secreted but also may increases the severity of ER stress and the rate of aggregate formation. Thus the heterotypic association of wild-type proSP-C with its mutant isoforms may also contribute to the development of lung diseases linked to BRISPC proteins.
In summary, the data from the present study show that expression of BRISPC isoforms elicits similar cellular reaction, suggesting activation of common cell signaling pathways in response to these misfolded proteins. The study has identified caspase-4 as one of the intermediate signals that activates caspase-3, thereby linking ER stress to apoptosis. In addition, through ER stress-independent and perhaps dysfunctional proteasome- and aggregate-dependent intermediates, cytochrome c is released from the mitochondria into the cytosol (Fig. 6). Moreover, induction of ER stress demonstrated in this study supports data from earlier reports showing that BRISPC expression resulted in the upregulation of ER stress response chaperone genes such as HDJ2/HSP40 and Bip/GRP78 (4, 5, 33). Investigation of the molecular mechanisms such as those described in this report will provide important information to understanding disease progression and may offer insights regarding specific targets for therapeutic intervention.
This work was supported by National Heart, Lung, and Blood Institute Grants HL-19737, HL074064, and P50 HL-56401 (to M. F. Beers) and HL-074064 Minority Supplement Program (to S. Mulugeta), American Lung Association Dalsemer Research Grant DA-188-N (to S. Mulugeta), and the Parker B. Francis Foundation Fellowship Grant (to S. Mulugeta).
We thank Dr. Masayuki Miura of the University of Tokyo (Tokyo, Japan) for the generous gift of the ERAI plasmids.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society