Expression of a Human Surfactant Protein C Mutation Associated with Interstitial Lung Disease Disrupts Lung Development in Transgenic Mice*

Surfactant Protein C (SP-C) is a secreted transmembrane protein that is exclusively expressed by alveolar type II epithelial cells of the lung. SP-C associates with surfactant lipids to reduce surface tension within the alveolus, maintaining lung volume at end expiration. Mutations in the gene encoding SP-C (SFTPC) have recently been linked to chronic lung disease in children and adults. The goal of this study was to determine whether a disease-linked mutation in SFTPC causes lung disease in transgenic mice. The SFTPC mutation, designated g.1728 G → A, results in the deletion of exon4, generating a truncated form of SP-C (SP-CΔexon4). cDNA encoding SP-CΔexon4 was constitutively expressed in type II epithelial cells of transgenic mice. Viable F0 transgene-positive mice were not generated after two separate rounds of pronuclear injections. Histological analysis of lung tissue harvested from embryonic day 17.5 F0 transgene-positive fetuses revealed that SP-CΔexon4 caused a dose-dependent disruption in branching morphogenesis of the lung associated with epithelial cell cytotoxicity. Transient expression of SP-CΔexon4 in isolated type II epithelial cells or HEK293 cells resulted in incomplete processing of the mutant proprotein, a dose-dependent increase in BiP transcription, trapping of the proprotein in the endoplasmic reticulum, and rapid degradation via a proteasome-dependent pathway. Taken together, these data suggest that the g.1728 G → A mutation causes misfolding of the SP-C proprotein with subsequent induction of the unfolded protein response and endoplasmic reticulum-associated degradation pathways ultimately resulting in disrupted lung morphogenesis.

Type II epithelial cells synthesize and secrete pulmonary surfactant, a complex mixture of phospholipids and proteins that coats the alveolar surface. Surfactant forms a bioactive film that effectively reduces the amount of work required during inspiration and prevents alveolar collapse at end expiration. The protein components of surfactant, in particular the lipophilic proteins surfactant protein B (SP-B) 1 and surfactant protein C (SP-C), facilitate the adsorption and spreading of lipids during the respiratory cycle and are critical for the formation and maintenance of the surfactant film. The importance of SP-C in mediating this process is underscored by the efficacy of exogenous surfactant preparations containing SP-C as the sole protein component (1)(2)(3).
SP-C is exclusively expressed in type II epithelial cells of the lung and is highly conserved among all species studied to date (4 -6). Human SP-C is synthesized as a 197-amino acid proprotein (molecular mass ϭ 21 kDa) consisting of a 35-amino acid mature peptide flanked by an N-terminal propeptide (residues 1-23) and a C-terminal domain (residues 59 -197). The intact proprotein traverses the regulated secretory pathway of the type II cell from the endoplasmic reticulum (ER) to the late endosome/multivesicular body where cleavage of the flanking N-and C-terminal domains liberates the extremely hydrophobic, bioactive mature peptide (7,8). Fusion of the multivesicular body with a lamellar body results in delivery of the mature peptide to the intracellular compartment in which the fully assembled surfactant complex is stored until it is secreted into the air space.
SP-C is an integral membrane protein that contains a single membrane-spanning domain located within the mature peptide (9). The newly synthesized proprotein is inserted into the ER membrane in a type II orientation with the N-terminal propeptide located in the cytosol and the C-terminal domain residing in the lumen of the ER (10,11). Trafficking of the proprotein through the regulated secretory pathway of the type II cell is dependent upon signals encoded in the N-terminal propeptide (11,12) and may be facilitated by oligomerization because the SP-C proprotein has been shown to form dimers and oligomers in transfected A549 cells (13). The C-terminal peptide is dispensable for trafficking and secretion of the mature peptide (11); however, the orientation of the C-terminal domain in the lumen of the ER subjects SP-C to scrutiny by ER quality control mechanisms.
Mutations in the human SP-C gene (SFTPC) have recently been linked to familial interstitial lung disease (ILD). Although numerous histopathological classifications exist for this diverse group of diseases, including nonspecific interstitial pneumonitis and usual interstitial pneumonitis, pulmonary fibrosis is generally regarded as the final common outcome. The index case was a patient diagnosed at 1 year of age with nonspecific interstitial pneumonitis (14). Sequence analysis of genomic DNA from the patient revealed a point mutation on one allele of SFTPC. The mother of the infant also harbored the mutation and was afflicted with lifelong lung disease. The mutation substitutes an adenosine for a guanosine at the first base pair of intron 4 (genomic DNA base 1728 (g.1728 G 3 A)), resulting in ablation of the normal donor splice site at the exon 4/intron 4 boundary (see Fig. 1A). Splicing of the mutant mRNA results in deletion of exon4, which encodes 37 amino acids in the C-terminal peptide of SP-C, ultimately resulting in the generation of a truncated proprotein of 17 kDa (SP-C ⌬exon4 ). The mutant proprotein was associated with a decrease in wild-type SP-C and a complete absence of mature SP-C in the BALF of the affected patient, suggesting that both SP-C ⌬exon4 and wildtype SP-C might be retained in the ER by the quality control apparatus.
A separate point mutation in the SFTPC locus (exon 5 ϩ 128 T 3 A) was also identified in an extended kindred with a history of ILD, including adults presenting with usual interstitial pneumonitis and children with cellular nonspecific interstitial pneumonitis (15). The mutation results in the substitution of a glutamine for a conserved leucine residue (L188Q) in the C-terminal domain of SP-C. Similar to patients with the g.1728 G 3 A mutation, affected individuals in this kindred carried the mutation on only one allele. Pedigree analysis demonstrated variability in the age of disease onset and phenotypic presentation, suggesting that multiple genetic and/or environmental factors were involved in the pathogenic process. To date, 11 mutations in the SFTPC locus have been linked to ILD, and all except one map to the C-terminal peptide of SP-C (16). This study was designed to test the hypothesis that a disease-linked mutation in the C-terminal peptide of SP-C causes lung disease.

DNA Constructs, Generation of Transgenic Mice, and Adenovirus
Production-Full-length human SP-C (SP-C  ) cDNA was cloned into pcDNA3 (Invitrogen) to generate SP-C 1-197 /pcDNA3. SP-C ⌬exon4 was generated by deleting nucleotides 325-435 (adenosine of start ATG is base pair 1) via overlapping polymerase chain reaction mutagenesis using SP-C 1-197 /pcDNA3 as a template and cloned into pcDNA3. A transgene construct was generated by subcloning SP-C ⌬exon4 into a puc19-based vector containing the 13-kb mouse SP-C promoter (17), rabbit globin intronic/exonic sequences, and a bovine growth hormone polyadenylation signal (Fig. 1A) as previously described (18). All completed constructs were sequenced bidirectionally to verify the integrity of the SP-C coding sequence. To generate transgenic mice, the transgene was excised from the vector DNA, purified, and microinjected into the male pronuclei of fertilized FVB/N oocytes by the University of Cincinnati Transgenic Core facility. Potential founder mice were identified by transgene-specific PCR analysis of tail DNA. Adenoviral constructs were generated by subcloning the SP-C  or SP-C ⌬exon4 construct from the pcDNA3 vector into the Adv2 shuttle vector (19). Recombination and adenovirus production were performed as described previously (19).
Histochemical and Western Analysis of Lung Tissue-Potential founder mice (F 0 ) were harvested by Caesarian section at E17.5. Left lung tissues were removed for Western blot analyses, and right lung tissues were fixed en bloc for light microscopy, immunohistochemistry, and in situ hybridization as previously described (20). For Western analysis, lung tissues were homogenized in phosphate-buffered saline containing 1% per volume protease inhibitor mixture (Sigma). Total protein concentration of the lung homogenate was determined by BCA assay (Pierce), and equal amounts of protein were separated by SDS-PAGE. Separated proteins were electrophoretically transferred to nitrocellulose membranes and probed with a polyclonal antibody directed against the N-terminal peptide of SP-C (21) or actin (a kind gift from Dr. James Lessard, Cincinnati Children's Hospital, Cincinnati, OH). Stripping of the antibody complexes was performed using Restore TM Western blot stripping buffer (Pierce). Immunohistochemistry was performed using rabbit polyclonal antibodies directed against the N-terminal propeptide of SP-C and Clara cell secretory protein (CCSP) (the latter antibody was kindly provided by Dr. Barry Stripp, University of Pittsburgh) at the indicated dilutions. Biotinylated secondary antibodies and a streptavidin-biotin-peroxidase detection system (Vector Laboratories, Inc.) were used to localize the antibody-antigen complexes in the tissues, as previously described (20). In situ hybridization was performed as previously described (5) using a 35 S-UTP-labeled, transgene-specific riboprobe directed against the bovine growth hormone polyadenylation signal (283-bp fragment).
Type II Cell Isolation, Adenoviral Infection, Metabolic Labeling, Immunoprecipitation, and in Vitro Transcription/Translation-Type II epithelial cells were isolated from SP-CϪ/Ϫ mice (22,23) using the method described by Rice et al. (24). 1 ϫ 10 6 cells/well were plated on 100% Matrigel (BD Pharmingen, San Diego, CA) in growth medium consisting of bronchial epithelial cell growth medium containing all accompanying additives except hydrocortisone (Clonetics, Walkersville, MD). The medium also included 10% charcoal-stripped fetal bovine serum (Sigma) and 10 ng/ml KGF (Peprotech, Rocky Hill, NJ). The cells were cultured at 37°C in a humidified incubator containing 5% CO 2 . Seventy-two hours post-isolation, the cells were infected with 50 multiplicity of infection of purified adenoviral particles encoding SP-C  or SP-C ⌬exon4 in infection medium containing 2% fetal bovine serum; infection medium was replaced with complete medium 90 min following infection. Forty-eight hours post-infection, the cells were metabolically labeled with 0.5 mCi/ml of [ 35 S]methionine/cysteine (ICN, Aurora, OH) for 4 h. The cell lysates were immunoprecipitated exactly as described previously (25) with 5 l of an antibody directed against the mature SP-C peptide (26). SDS-PAGE and autoradiography was performed as previously described (25). SP-C 1-197 /pcDNA3 and SP-C ⌬exon4 /pcDNA3 were transcribed and translated in vitro in the presence of [ 35 S]methionine/cysteine (ICN) using the TNT® coupled reticulocyte lysate system (Promega, Madison, WI). The completed reactions were analyzed by SDS-PAGE/autoradiography.
HEK293 Cell Culture and Transfection-HEK293 cells were purchased from ATCC (Manassas, VA). Growth medium consisted of Richter's medium (Biowhittaker, Walkersville, MD) containing 10% fetal bovine serum (Sigma), 2 mM L-glutamine and 1 unit/ml penicillin and streptomycin (Sigma). The cells were cultured at 37°C in a humidified incubator containing 5% CO 2 . For proteasome inhibitor experiments, 2 ϫ 10 5 cells were plated into a 12-well plate 24 h prior to transfection. The cells were transiently transfected with 1 g/well SP-C 1-197 /pcDNA3 or SP-C ⌬exon4 /pcDNA3 using LT-1 reagent (Mirus, Madison, WI). Fourhours prior to harvest, the cells were treated with 5 M MG-132 (Calbiochem, La Jolla, CA) or Me 2 SO as a vehicle control. The cells were harvested in phosphate-buffered saline, sonicated immediately, the total protein content was assessed by BCA assay, and equal amounts of protein were analyzed by SDS-PAGE/Western blotting with proSP-C or actin antisera as described above.
Immunofluorescence-HEK293 cells transfected with SP-C ⌬exon4 or SP-C  were plated on poly-D-lysine coated coverslips. 4 h prior to fixation, the cells were treated with 5 M MG-132 or Me 2 SO vehicle control (Calbiochem, La Jolla, CA) as indicated. The cells were fixed with 4% paraformaldehyde in phosphate-buffered saline, permeabilized with 0.1% Triton X-100, and stained with a polyclonal antibody directed against the N-terminal propeptide of SP-C for 2 h at 37°C. The cells were washed and incubated with anti-rabbit, fluorescein isothiocyanate-conjugated secondary antibody (Jackson Immunoresearch, West Grove, PA) for 2 h at 37°C. The cells were washed three times with phosphate-buffered saline, washed once with distilled H 2 O, and mounted on slides with Vectashield mounting medium (Vector Laboratories, Inc.). Fluorescence was imaged with a Nikon microscope using a UV lamp and a fluorescein isothiocyanate filter. The images were captured using an Optronics MagnaFire digital color camera.
BiP/Luciferase Assays-The BiP/Luciferase reporter (a kind gift from Dr. Randal Kaufman, University of Michigan) consists of a minimal BiP promoter (nucleotides Ϫ457 to ϩ33) containing the ER stress element placed immediately upstream of the luciferase gene in the pGL3-basic plasmid (Promega) (27). The ␤-galactosidase plasmid, pSV-␤-galactosidase, was purchased from Promega. 2 ϫ 10 5 HEK293 cells were plated in a 12-well dish 24 h prior to transfection. The cells were co-transfected with three plasmids including: 1) 75 or 250 ng of one of the test plasmids (SP-C  , SP-C ⌬exon4 , or pcDNA3 as an empty vector control), 2) 250 ng of BiP/luciferase reporter plasmid, and 3) 75 ng of ␤-galactosidase plasmid. The cells were harvested 48 h post-transfection in Glo Lysis Buffer (Promega). The luciferase activity was measured with a Bright Glo™ luciferase assay system (Promega), and the ␤-galactosidase activity was measured using the Luminescent ␤-galactosidase detection kit II (BD Biosciences, Palo Alto, CA) in a luminometer. The data are plotted as luciferase activity/␤-galactosidase activity to correct for transfection efficiency among samples. The presented data represents one of three independent transfection experiments with each group performed in triplicate. Statistical differences between groups were assessed by one-way analysis of variance.

Generation of SP-C ⌬exon4
Transgenic Mice-To test the hypothesis that mutations in the C-terminal peptide of SP-C cause lung disease, SP-C ⌬exon4 was expressed in type II epithelial cells of transgenic mice. The SP-C ⌬exon4 construct was synthesized by deleting the nucleotides encoding the amino acids of exon4 (residues109 -145) via overlapping PCR mutagenesis using human SP-C cDNA as a template. Expression of SP-C ⌬exon4 was specifically targeted to type II epithelial cells of the lung using the 13-kb mouse SP-C promoter (Fig. 1B). Despite two rounds of pronuclear injections into FVB/N eggs, no transgene-positive progeny were recovered, suggesting that expression of SP-C ⌬exon4 was associated with neonatal lethality. The construct was therefore injected again, and the lungs were harvested from potential founder mice (F 0 ) at E17.5 for histochemical and biochemical analyses. Thirty embryos were recovered from this injection, five of which had integrated the transgene into the genome.
SP-C ⌬exon4 Disrupts Lung Organogenesis in Transgenic Mice-Two of five transgene-positive F 0 fetuses displayed disrupted lung organogenesis, as assessed by hematoxylin and eosin staining of fetal lung tissue (Fig. 2), whereas the remaining three F 0 mice were morphologically indistinguishable from weight-matched, wild-type fetuses. Lung tissue from the most severely affected animal (TG#1) was extremely hypoplastic and characterized by large cystic saccules, little branching morphogenesis, and loss of typical distal acinar structures ( Fig. 2A). Although the lungs of the less affected animal (TG#2) were comparable in size to wild-type fetuses, branching morphogenesis was also disrupted, albeit to a lesser extent than that observed in TG#1 (Fig. 2B). The lung structure of the other three F 0 mice (one of which is depicted in Fig. 2C as TG#3) was completely normal.
Analysis of the hematoxylin and eosin-stained sections at higher magnification revealed the presence of vacuolated (Fig.   3A, arrows) and hypertrophic (Fig. 3A, arrowheads) epithelial cells in TG#1. The epithelium of TG#2 appeared largely intact and contained few vacuolated cells (Fig. 3B). Both TG#1 and TG#2 contained a significant amount of debris in the proximal and distal airways. The airway debris was more prominent in TG#2 than in TG#1 and consisted of sloughed epithelial cells and macrophages recruited into the air spaces. In contrast, the epithelium in TG#3 and WT were intact, devoid of airway debris, and supported by condensed mesenchymal tissue indicative of the canalicular stage of murine lung development (Fig.  3, C and D). Collectively these data demonstrate that expression of SP-C ⌬exon4 in type II epithelial cells of transgenic mice disrupted normal lung organogenesis, ultimately leading to neonatal lethality. The prominent abnormalities associated with this disruption included cytotoxicity, hypoplasticity, and perturbation of branching morphogenesis.
To determine whether variability in phenotype was associated with the level of SP-C ⌬exon4 expression, in situ hybridization was performed with a radiolabeled riboprobe specific for transgenic mRNA. A gradient of SP-C ⌬exon4 mRNA expression was detected among F 0 mice with the highest expression observed in TG#1 and the lowest expression observed in TG#3 Hematoxylin and eosin stained fetal lung sections from mice expressing SP-C ⌬exon4 (E17.5) are shown. TG#1 represents a section through all right lobes of the lung compared with a section through one right lobe for the remaining animals. Note severe hypoplasia in TG#1 and presence of large, cystic saccules and a loss of mesenchymal tissue in TG#1 and TG#2 as compared with the wild-type control (WT). TG#3 was morphologically indistinguishable from the wild-type control. Scale bar, 500 m. (Fig. 4, A-C). The antisense signal in WT (Fig. 4D) was indistinguishable from sense controls (data not shown). These results demonstrate that the expression level of SP-C ⌬exon4 was correlated with the severity of the lung phenotype. The expression pattern of SP-C ⌬exon4 was assessed by immunohistochemistry using an antibody specific for the N-terminal propeptide of SP-C (proSP-C) (Fig. 4, E-H). Because this antibody detects both endogenous SP-C and SP-C ⌬exon4 , the primary antibody was titered to a concentration at which only SP-C ⌬exon4 was detected (1:44,000); endogenous SP-C staining was detected at a primary antibody dilution of 1:1,000 but not 1:4,000. Intense SP-C immunoreactivity was observed in the epithelial cells lining the cystic saccules of TG#1 (Fig. 4E), confirming successful targeting of SP-C ⌬exon4 to the distal epithelium. In addition, high expression of SP-C ⌬exon4 in TG#1 was associated with sloughing of the epithelium, leading to an accumulation of SP-C-positive cellular debris in the air spaces (Fig. 4E, inset). Interestingly, SP-C immunoreactivity was only detected in the air spaces of TG#2 and not in the epithelium at the 1:44,000 dilution, suggesting that intracellular SP-C ⌬exon4 protein was turned over rapidly in this animal ( Fig. 4F and inset). Lung tissue from TG#3 was completely devoid of SP-C staining at the 1:44,000 dilution (Fig. 4G), even though SP-C ⌬exon4 mRNA was detected in this mouse (Fig. 4C). For comparison, the endogenous SP-C staining pattern of distal epithelial cells is shown on wild-type tissue at a 1:1,000 dilution of primary antibody (Fig.  4H). The staining pattern for a proximal epithelial cell marker, CCSP, appeared normal, suggesting that cell specification was not perturbed even in the most severely affected lungs (Fig. 4, I-L). Taken together, these data are consistent with the hypothesis that SP-C ⌬exon4 exerts a dose-dependent, cytotoxic effect in the respiratory epithelium of transgenic mice.
To determine the relative expression levels of SP-C ⌬exon4 protein in transgenic F 0 mice, Western blot analysis was performed on lung tissue using the proSP-C antibody. SP-C proprotein is not normally detected by Western blotting because of rapid processing to the mature peptide in the biosynthetic pathway of the type II cell (Fig. 5A, WT). Immunoreactive SP-C was readily detected in TG#1 (Fig. 5A), and the size of the proSP-C positive band (M r ϭ 17,000) corresponded to the predicted molecular weight of SP-C ⌬exon4 . Immunoreactive proSP-C also co-migrated with newly synthesized SP-C ⌬exon4 transcribed and translated in vitro from a mammalian expression vector (Fig. 5B). These data demonstrate that the expression of SP-C ⌬exon4 in TG#1 detected by Western analysis correlates with the expression levels observed by immunohistochemistry. Interestingly, SP-C ⌬exon4 protein was undetectable in the lung homogenate of TG#2, despite abnormal lung morphogenesis, suggesting that the mutant protein was rapidly degraded in this animal.
SP-C ⌬exon4 Is Rapidly Degraded in Vitro-Newly synthesized SP-C is a type II integral membrane protein in which the C-terminal peptide domain resides in the lumen of the ER (10,  11). Although the function of the C-terminal peptide is unknown, mutations in this region may result in misfolding of the protein, resulting in retention in the ER and incomplete processing of the proprotein. To determine whether deletion of exon4 prevented processing of the proprotein, type II epithelial cells were isolated from SP-CϪ/Ϫ mice infected with adenoviral particles encoding SP-C  or SP-C ⌬exon4 and cell lysates immunoprecipitated with an antibody directed against the mature SP-C peptide. In type II cells infected with SP-C 1-197 , the full-length proprotein (M r ϭ 21,000), two processing intermediates and mature SP-C (M r ϭ 4,000) were detected (Fig. 6,  lanes 1 and 2). An identical banding pattern was observed when endogenous SP-C was immunoprecipitated from metabolically labeled type II cells isolated from a wild-type mouse (data not shown). In contrast, only the mutant proprotein (M r ϭ 17,000) and a smaller immunoreactive form were detected in cells expressing SP-C ⌬exon4 (Fig. 6, lanes 3 and 4). These results show that SP-C ⌬exon4 was not completely processed to the mature peptide, suggesting that the mutant proprotein was not sorted to the distal compartments of the secretory pathway in SP-CϪ/Ϫ type II cells.
To determine whether SP-C ⌬exon4 was degraded early in the biosynthetic pathway, SP-C  or SP-C ⌬exon4 was transiently transfected into HEK 293 cells in the presence or absence of the proteasome inhibitor MG-132. The cell lysates were harvested 4 h after the addition of MG-132, and Western analysis was performed with the proSP-C antibody. Robust expression of SP-C  was detected in the absence of proteasome inhibitor (Fig. 7, lanes 3 and 4); in contrast, only faint immunoreactive bands, corresponding to the mutant proprotein, were detected in cells expressing SP-C ⌬exon4 , demonstrating that the mutant protein was rapidly degraded in the absence of the proteasome inhibitor (Fig. 7, lanes 7 and 8). SP-C ⌬exon4 mutant proprotein was readily detected following MG-132 treatment and approached levels observed in cells expressing SP-C   (Fig. 7,  lanes 5 and 6). Similarly, expression of SP-C ⌬exon4 in transiently transfected HeLa or 3T3 cells was only detectable in the presence of MG-132, demonstrating that ER-associated degradation (ERAD)-dependent turnover of the mutant proprotein was not cell type-specific (data not shown). MG-132 had no effect on the level of SP-C  , indicating that little or no wild-type proprotein was degraded via the proteasome pathway (Fig. 7, lanes 1 and 2).
To confirm that SP-C ⌬exon4 was not exported from the ER, immunolocalization of SP-C was performed on transfected HEK293 cells expressing either SP-C  or SP-C ⌬exon4 . SP-C 1-197 exhibited a punctuate staining pattern that was unaltered by proteasome inhibition (Fig. 8, top panels). This staining pattern is consistent with trafficking of the wild-type protein to the lysosome, the compartment to which regulated secretory proteins, such as SP-C, traffic in a cell that lacks such a pathway (28). In contrast, low levels of protein were detected in cells that expressed SP-C ⌬exon4 in the absence of a proteasome inhibitor, consistent with results obtained by Western analysis (Fig. 8, bottom panels). Treatment with MG-132 prevented degradation of the SP-C ⌬exon4 , revealing a diffuse staining pattern indicative of ER localization (Fig. 8, bottom panels). These results support the hypothesis that SP-C ⌬exon4 is recognized as a misfolded peptide, retained in the ER, and rapidly degraded via the ERAD pathway in a non-type II epithelial cell line.
SP-C ⌬exon4 Induces ER Stress-BiP is an abundant chaperone protein whose primary function is to facilitate the folding of proteins in the ER. Transcription of BiP is increased in response to the accumulation of unfolded or misfolded proteins in the ER and thus serves as a classical marker for the induction of ER stress pathways (29). To determine whether SP-C ⌬exon4 induced ER stress, mammalian expression vectors encoding SP-C ⌬exon4 , SP-C  , or an empty vector (pcDNA3) were individually transfected into HEK293 cells with a reporter vector consisting of a minimal BiP promoter driving luciferase (BiP/ Luc) (Fig. 8). The cell lysates were harvested 48 h post-transfection and analyzed for luciferase activity and SP-C levels via Western analysis. The cells transfected with BiP/Luc and subjected to a 6-h exposure of 10 g/ml tunicamycin showed a 2.4-fold increase in luciferase activity over the pcDNA3 control, indicating that the BiP promoter was indeed responsive to a known ER stress-inducing agent (Fig. 9A). 75 ng of SP-C ⌬exon4 co-transfected with the BiP/Luc reporter resulted in a 2-fold increase in luciferase activity compared with cells transfected with the empty vector control. Co-transfection of an equivalent amount of SP-C  and BiP/Luc caused a modest increase in luciferase activity compared with pcDNA3 but failed to reach significance (Fig. 9A). When the amount of input cDNA was increased to 250 ng, SP-C ⌬exon4 augmented luciferase activity FIG. 6. Inappropriate processing of SP-C ⌬exon4 in SP-C؊/؊ type II cells. Type II cells were isolated from SP-CϪ/Ϫ mice and cultured on 100% Matrigel for 48 h. The cells were infected with adenoviral particles encoding SP-C   (duplicates, lanes 1 and 2) or SP-C ⌬exon4 (duplicates, lanes 3 and 4). Forty-eight hours post-infection, the cells were metabolically labeled with [ 35 S]cysteine/methionine, and the cell lysates were immunoprecipitated with a polyclonal antibody directed against the mature SP-C peptide. The immunoprecipitates were separated by SDS-PAGE and subjected to autoradiography. The top portion of the gel containing proproteins was exposed to film for 24 h, and the bottom portion was exposed for 72 h. The molecular mass markers are indicated on the left.  1-4) or SP-C ⌬exon4 (lanes 5-8) were transiently transfected into HEK293 cells. Twenty-four hours post-transfection, the proteasome inhibitor MG-132 was added (lanes 1, 2, 5, and 6) for 4 h prior to harvest. The cell lysates were separated by SDS-PAGE and immunoblotted with an antibody specific for the Nterminal propeptide of SP-C. The blot was stripped and reprobed with an anti-actin antibody for a loading control. The presented data represent three experiments. The molecular mass markers are indicated on the left.
3.5-fold, whereas SP-C 1-197 increased luciferase activity 2.4fold over that observed for pcDNA3 (Fig. 9A). The increases in luciferase activity observed with SP-C ⌬exon4 were statistically significant compared with both SP-C  and pcDNA3 for the two input quantities of cDNA tested. Western analyses on cell lysates from the 250-ng input group showed high expression of SP-C   (Fig. 9B, lanes 1-3). SP-C ⌬exon4 was undetectable in the cell lysates, despite the increase in BiP promoter activity (Fig. 9B, lanes 4 -6) consistent with rapid degradation of the mutant proprotein. Collectively, these results indicate that the expression of SP-C ⌬exon4 in HEK293 cells elicits an ER stress response in a dose-dependent manner. DISCUSSION Mutations in the gene encoding human surfactant protein C are associated with chronic lung disease in both children and adults. The goal of this study was to determine whether the g.1728 G 3 A (SP-C ⌬exon4 ) point mutation in the SFTPC locus was directly linked to the pathogenesis of lung disease. This hypothesis was tested by generating transgenic mice that expressed SP-C ⌬exon4 in type II cells of the respiratory epithelium. SP-C ⌬exon4 caused a dose-dependent perturbation of lung development associated with epithelial cell cytotoxicity. Transient expression of SP-C ⌬exon4 in isolated type II epithelial cells or HEK293 cells resulted in incomplete processing of the proprotein, a dose-dependent increase in BiP transcription, trapping of the proprotein in the ER, and rapid degradation via a proteasome-dependent pathway. Taken together these data suggest that the g.1728 G 3 A mutation leads to misfolding of the SP-C proprotein with subsequent induction of unfolded protein response (UPR) and ERAD pathways.
Lung development was profoundly disrupted despite the fact that expression of the SP-C ⌬exon4 protein was restricted to one cell type and occurred in the presence of two wild-type alleles. Three lines of evidence implicate SP-C-mediated cytotoxicity as the basis for altered lung structure. First, epithelial cells expressing high levels of the transgene exhibited cell swelling consistent with necrosis. Second, the sloughed respiratory epithelium and the cellular debris detected in the airways of two independent F 0 animals stained intensely for proSP-C at an antibody dilution that detected only the transgene-derived protein. Third, macrophage infiltrates were present in the lungs of both affected animals. Because macrophages are never observed in the fetal lung in the absence of inflammation, it is likely that these cells were recruited to the lung following cell injury.
To determine whether SP-C ⌬exon4 -induced dysmorphogenesis was linked to altered epithelial cell specification, markers of the proximal and distal respiratory epithelium were analyzed in lung tissues from three F 0 animals. The staining patterns of the proximal epithelial cell marker CCSP and the distal epithelial cell marker proSP-C were normal, suggesting that cell specification was not altered. Therefore, it is likely that inappropriate epithelial cell death resulted in disruption of branching morphogenesis rather than a defect in cell specification. It is well established that epithelial-mesenchymal in-   , SP-C ⌬exon4 , or an empty vector control (pcDNA3) were individually transfected into HEK293 cells with a reporter vector consisting of a minimal BiP promoter driving the firefly luciferase gene (BiP/Luc). A third plasmid encoding ␤-galactosidase was co-transfected to standardize the samples for transfection efficiency. The amount of input cDNA for the tested plasmids was either 75 ng (top panel) or 250 ng (bottom panel), whereas the inputs of the BiP/Luc and ␤-galactosidase plasmids were constant. The cells were harvested 48 h post-transfection, and the cell lysates were analyzed for luciferase and ␤-galactosidase activity using a luminometer. The cells transfected with the empty vector control were subjected to a six-hour exposure of 10 g/ml tunicamycin (pcDNA3 ϩ TM) prior to harvest to demonstrate the responsiveness of the BiP/Luc reporter to a known ER stress-inducing agent. *, p Ͻ 0.001 versus pcDNA3; #, p Ͻ 0.001 versus SP-C  . B, the cell lysates from A (lower panel) were separated by SDS-PAGE and immunoblotted with an antibody directed against the N-terminal propeptide of SP-C. Lanes 1-3 represent triplicates for SP-C  , and lanes 4 -6 represent triplicates for SP-C ⌬exon4 . The data are expressed as the means Ϯ S.D.
teractions are absolutely required for proper branching morphogenesis in numerous organs including the lung (30). Ablation of type II epithelial cells would effectively terminate signaling between the two cell compartments, resulting in altered morphogenetic signaling. It is unlikely that the dysmorphogenesis was solely due to overexpression of transgene protein in type II epithelial cells because mice expressing SP-B ⌬C or lysozyme transgenes were viable with no lung abnormalities (31,32). A phenotype similar to that seen in the SP-C ⌬exon4 mice was observed in mice expressing high levels of the SP-C mature peptide, SP-C 24 -58 , or diptheria toxin in type II cells (33,34). Collectively, these data suggest that the demise of fetal type II epithelial cells, irrespective of the causal insult, leads to altered lung morphogenesis.
Deletion of exon4 from SP-C resulted in incomplete processing of the mutant proprotein in isolated type II epithelial cells. Processing of the N-terminal propeptide and the C-terminal peptide occurs in the multivesicular body of the type II cell leading to the generation of the 4-kDa active peptide (7,8). Lack of SP-C ⌬exon4 processing suggested that the mutant proprotein did not traffic to the multivesicular body and was retained in a proximal compartment of the secretory pathway; confocal microscopy identified this compartment as the ER. Quality control mechanisms within the ER ensure the correct folding and assembly of polypeptides prior to export from this compartment. Accumulation of unfolded or misfolded protein triggers an ER-to-nucleus signal transduction pathway, the UPR, which up-regulates the production of chaperone proteins, such as BiP, within the ER. Failure to fold under these conditions results in induction of ERAD leading to retrotranslocation of the terminally misfolded protein and degradation by the ubiquitin-proteasome pathway (for review see Refs. 29 and 35). Expression of SP-C ⌬exon4 in HEK293 cells induced a dose-dependent increase in BiP transcription and rapid degradation of the proprotein via proteasome-dependent mechanisms. Proteasome-dependent degradation of SP-C ⌬exon4 in HEK293 cells together with the inability to detect SP-C ⌬exon4 protein in TG#2 and incomplete processing of the mutant proprotein in type II cells support the hypothesis that SP-C ⌬exon4 is recognized as a misfolded protein within the ER and rapidly degraded via ERAD. Wild-type SP-C 1-197 also caused an increase in BiP transcription at the higher input dose of cDNA; however, unlike SP-C ⌬exon4 , SP-C  was successfully exported from the ER. The increase in BiP transcription was therefore most likely due to robust expression of wild-type SP-C resulting in an increase in unfolded substrate that triggered the UPR. Collectively these results indicate that SP-C ⌬exon4 induces ERAD in HEK293 cells resulting in selective degradation of mutant but not wild-type SP-C.
The high level of SP-C ⌬exon4 expression in transgenic mice may have been sufficient to saturate ERAD leading to epithelial cell death and disruption of lung morphogenesis. Lower levels of mutant SP-C proprotein may cause a milder phenotype leading to postnatal ILD observed in human patients; this hypothesis remains to be tested. Attenuation of translation, which accompanies induction of UPR, may also have contributed to dysmorphogenesis by inhibiting new protein synthesis during a critical stage of lung growth and differentiation. We also cannot dismiss the possibility that the loss of SP-C in the air spaces contributed to the severity of the disease in humans and transgenic mice. ILD was detected in a family with no detectable SP-C in BALF as well as in SP-CϪ/Ϫ mice (22,36). However, SP-C deficiency cannot be the sole cause of disease in transgenic mice because lung structure and function is normal in newborn SP-CϪ/Ϫ mice (23).
A previous study in transiently transfected A549 cells demonstrated that deletion of exon 4 in the context of SP-C/GFP fusion proteins resulted in ubiquitination and aggresome formation (37). The formation of aggresomes suggested that mutant SP-C fusion protein had a prolonged half-life and was resistant to degradation. The proteasome-dependent turnover of SP-C ⌬exon4 in HEK293 cells is consistent with ubiquitination of the SP-C cysteine mutants in A549 cells. However, rapid turnover of SP-C ⌬exon4 both in vitro and in vivo is inconsistent with aggresome formation. Although ultrastructural analysis was not performed on SP-C ⌬exon4 F 0 mice because of the limited amount of lung tissue, aggresomes were not detected in transgenic mice expressing SP-C 24 -58 , which displayed a similar phenotype to SP-C ⌬exon4 transgenic mice (33). Perhaps very high expression of the SP-C/GFP mutant in a few isolated A549 cells overwhelmed the degradative capacity of the proteasome, leading to aggresome formation in a subset of cells. The frequency of aggresome formation was not reported in the study by Wang et al. (37); however, in HEK293 cells transiently expressing high levels of the folding mutant CFTR ⌬F508 , only 5-15% of cells contained aggresomes (38). Taken together, we postulate that the constitutive expression of misfolded SP-C ⌬exon4 overwhelmed the degradative capacity of ERAD machinery, resulting in chronic induction of ER stress pathways, type II cell injury, and disrupted lung morphogenesis.