Main

Pulmonary surfactant is a biochemically heterogeneous surface-active mixture of phospholipids and proteins secreted into the alveolar space by type II epithelial cells (1). Formation of the surfactant film at the air-liquid interface assures alveolar stability by reducing surface tension along the epithelial lining. It is now well accepted that dysfunction of the surfactant system plays an important role in the development or progression of both congenital and acquired lung diseases (1). Respiratory distress syndrome of the premature newborn due to a relative surfactant deficiency from either genetic or developmental regulatory disruption of the pulmonary surfactant production is the best-known clinical manifestation. In adults, abnormalities in surfactant component expression, metabolism, or functional inactivation have all been observed, most frequently in the context of acute lung injury (e.g. acute respiratory distress syndrome) (2). However, absolute and functional surfactant defects have been reported in other acute and subacute lung injuries including systemic sepsis, asthma, and bleomycin lung toxicity (3–5). In addition, disorders of surfactant homeostasis that result in an overabundance of alveolar surfactant material (PAP) have been associated with a variety of specific clinical entities including endogenous anti-GM-CSF antibodies (primary PAP), myeloproliferative disease, inorganic dusts, or select infectious agents such as Pneumocystis carinii (6).

Analysis of surfactant has identified four unique proteins, designated SP-A, -B, -C, and -D (reviewed in refs. 7 and 8). The hydrophilic proteins, SP-A and SP-D, are members of the collectin family of C-type lectins that include a number of molecules with known host-defense functions (7). The hydrophobic surfactant proteins, SP-B and SP-C, augment adsorption, spreading, and surface tension properties of phospholipid films (8). Surfactant protein C is synthesized as a proprotein of Mr 21,000 that undergoes integral membrane association and posttranslational removal of NH2 and COOH propeptides to yield a biophysically active form (9). The posttranslational processing of a protein integrally associated with intracellular membranes before exocytosis of the membrane anchor represents a novel mechanism in mammals for the production of a secreted protein (10–13).

Generation of an SP-C null mouse has been reported (14). The phenotype of the knockout is viable with apparently normal lung function at early ages most likely due to redundancy in biophysical activity afforded by the presence of a second hydrophobic surfactant protein, SP-B. However, when backcrossed onto a different strain, SP-C knockout mice develop lung inflammation and parenchymal remodeling in adulthood suggesting genetic background may be a modifier of an SP-C deficient state (15).

Recently, there have been two full-length reports of two different mutations in the human SP-C gene associated with the production of abnormal forms of proSP-C, an absence of mature SP-C, and chronic interstitial pneumonias presenting as nonspecific interstitial pneumonia (NSIP) in affected children and usual interstitial pneumonia (UIP) in surviving adults (16,17). In both circumstances, the mutation mapped to the distal portion of the COOH flanking propeptide and was expressed in a heterozygous fashion. This region of proSP-C (F94-I197) has recently been linked with several unrelated proteins all containing a novel sequence domain termed Bri (BRICHOS) found in degenerative and proliferative disorders in other organ systems including Bri3, associated with chronic neurodegenerative dementia in British and Danish kindreds, and CA11 found in chondrosarcomas (18).

We have shown that expression of one of these BRICHOS SP-C mutations (ΔExon4) in epithelial cells in vitro yields to the formation of intracellular protein aggregates and diversion of both wild-type and mutant forms to a novel intracellular compartment, the aggresome (19). These findings, along with recent work showing that expression of the SP-CΔExon4 mutation in transgenic mice disrupts lung morphogenesis (20), suggest that lung injury from expression of proSP-C proteins containing mutations in the BRICHOS domain of the propeptide is due to both a toxic gain of function from a cellular response to an aberrant, aggregating integral membrane protein as well as a dominant negative affect that down-regulates alveolar SP-C content.

In this report, we characterize the phenotype of a patient with a spontaneous mutation in a region in proSP-C spatially distinct from the aforementioned BRICHOS mutations. The heterozygous missense substitution of lysine for glutamic acid at position 66 of the proximal COOH flanking propeptide (hSP-CE66K) was associated not only with the presence of interstitial lung disease (NSIP) in the affected patient, but also with pulmonary alveolar proteinosis. Furthermore, in vitro evaluation of trafficking of this abnormal proSP-C isoform demonstrates that this mutant was targeted abnormally to early endosomal compartments that radically differ from the pattern observed for previously reported SP-C mutations. Taken together, in vitro data coupled with the clinical findings suggest that this mutant form of SP-C protein hSP-CE66K represents a new class of mutant proSPC forms that accumulate outside the normal biosynthetic pathway and are capable is inducing cell damage thus proving to be equally or even more harmful than a total absence of SP-C expression.

METHODS

Index Patient Summary

The index patient is a full-term male product of nonconsanguineous parents (40 wk gestation). The pregnancy and delivery were uneventful with a recorded birth weight of 3900 g. The infant exhibited no immediate perinatal problems and was discharged to home on hospital d 4. On the d 13 of life he was admitted to a local hospital with tachypnea, cyanosis, and hypoxemia.

Initial workup up of metabolic/infectious/immunologic etiologies was negative and on hospital d 3 the patient was transferred to a university hospital level-3 pediatric intensive care unit. Chest x-rays showed a persistence of diffuse bilateral pulmonary infiltrates in the lungs. The infant underwent BAL with a 3.5-mm flexible bronchoscope in the main bronchi in which approximately 2 mL/kg of saline were introduced and immediately withdrawn three times. Typical recovery was approximately 80% of the instilled volume. Analysis of this BAL and two additional samples, collected at 2-wk intervals, revealed increased cellularity containing foamy-appearing mononuclear cells but was persistently negative for bacterial or viral pathogens.

At 5 wk of age, ultrasound demonstrated the presence of pyloric hypertrophy with stenosis and corrective surgery was performed. The patient remained O2 dependent and an open lung biopsy was performed at the age of 2 mo that revealed the presence of NSIP (Fig. 1). The baby was eventually discharged to home requiring both mild ventilatory support (after tracheotomy) and supplemental O2.

Figure 1
figure 1

Photomicrographs of lung tissue from index patient. (a, b) Low and higher power views of hematoxylin and eosin–stained section of open lung biopsy specimen obtained at 2 mo of age demonstrating a marked thickening of alveolar septa due to spindle-shaped cells, a mild interstitial chronic inflammation, and a hyperplasia of type II pneumocytes (NSIP). (c) Immunostaining (red staining) for vimentin indicated that thickening of alveolar septa was mainly due to mesenchymal cells. Bar = 25 μm.

Full size image

Analysis of patient pathologic material and DNA was approved by Institutional Review Boards at the University of Pennsylvania, University of Padua, Italy, and Institute of Pathology, Bochum Germany.

Surfactant protein antisera.

Monospecific polyclonal antibodies against human SP-A and SP-B were produced for use in both immunoblotting and ELISA analysis. Antiserum against SP-A was made by injection of purified surfactant SP-A obtained from lavage of a patient with alveolar proteinosis. SP-A, triple purified by butanol precipitation and extraction with 5 mM Tris HCl, chromatography on a column of mannose-agarose, and by two-dimension PAGE was mixed with Freund's complete adjuvant and administered subcutaneously to rabbits. The antiserum, obtained after three boosting injections of SP-A in incomplete Freund's adjuvant, does not react against SP-D.

Polyclonal antiserum against SP-B was made in an identical fashion using SP-B purified from human PAP material surfactant using Sephadex LH-60 (Amersham Biosciences, Little Chalfont, Buckinghamshire, UK) eluted with choloform:methanol:0.1 N HCl (47.5:47.5:5). The antiserum, obtained after initial injection in Freund's adjuvant and three booster injections of SP-B in incomplete Freund's adjuvant, recognizes SP-B but not SP-C.

For immunocytochemical studies, a previously described monospecific polyclonal SP-B antiserum was used (21). Monospecific polyclonal antiserum against mature human SP-C was kindly provided by Altana Pharmaceuticals, Inc., Konstanz, Germany, and has been previously characterized (22,23). Polyclonal anti SP-D was obtained commercially from Chemicon, Inc., Temecula, CA. Monospecific polyclonal proSP-C antisera against NH2 terminus and COOH terminal propeptide domains were produced from synthetic peptide antigens and have been previously characterized (10). Each antiserum recognizes human proSP-C21 but does not recognize mature SP-C.

Marker antisera.

Anti-CD63 antiserum was purchased from Immunotech, Inc. (Marseille Cedex, France). A MAb against EEA-1 was purchased from BD Transduction Laboratories, Inc. (San Diego, CA). Anti-ubiquitin MAb was purchased from Chemicon, Inc. Texas Red (TR)-conjugated monoclonal and polyclonal antibodies were obtained from Jackson Immunoresearch Laboratories, Inc. (West Grove, PA).

PEGFP-C1 plasmid was purchased from Clontech, Inc. (Palo Alto, CA).

Preparation and Analysis of DNA

Patient and parental genomic DNA was purified from whole blood using a commercially available kit (Pure Gene, Gentra Systems, Minneapolis, MN).

Three amplicons corresponding to the numbering in the published sequence of SP-C (24–26) spanning exon 1 (genomic positions 596–662), exons 2 to 3 (genomic positions 1366–1994), and exons 4 to 5 (genomic positions 2228–2851) were amplified by PCR using the following primer sets: exon 1: 223–753: forward: 5′ ACC CTT GGC CAC CTG AAA GTT C 3′; reverse: 5′ AAG GAA ACA GGC CAG GGA GGA TAG 3′; exons 2 and 3: 1294–2078: forward: 5′AGC TTG TAT AGG GAG AAG AG 3′; reverse: 5′ GGA TGT GGA TGA AGT GGC 3′; exons 4 and 5: 2147–2891: forward: 5′ TAG AGG GAG GTG GCT AAG 3′; reverse: 5′ AGC GCT CCT CCG CTG CTG ATC 3′. PCR reactions containing 1 μg genomic DNA per 100 μL and 0.5 μM forward and reverse reaction primers, 200 μM dNTP mixture, and 0.5 μL of Taq polymerase (5 units/μL) were performed under the following conditions: denaturation at 95°C for 5 min followed by 30 cycles of denaturation at 95°C for 40 s, annealing 60°C for 40 s, and elongation at 72°C for 90 s.

Purified amplicons were subjected to automated sequencing in both directions by the Core Sequencing Facility at the University of Pennsylvania and compared with the published sequences (24–26).

Analysis of BAL

Biochemical analysis of BAL.

Protein concentration was determined by the method of Polacheck et al. (27). After organic extraction, total phospholipid content was measured by the method of Bartlett (28). The phospholipid composition of extracted lipids from BAL was determined by two-dimensional thin-layer chromatography according to Poorthuis et al. (29) and subsequent densitometry with a GS 700 Imaging Densitometer, (Bio-Rad, Huntingdon Station, NY). The fatty acid profile of PC was determined as follows: after preparative two-dimensional thin-layer chromatography, PC eluted from the silica was transmethylated and fatty acids were analyzed by gas chromatography using a PerkinElmer 8320 gas chromatograph (PerkinElmer Instruments, Norwalk, CT), equipped with a 30-m capillary column (0.32 mm internal diameter, Omegawax 320, Supelco Inc., Bellefonte, PA).

To measure SP-A, an antigen competition ELISA (range of detection 0–25 ng) was performed on BAL as previously published (30). The lower limit of detection of SP-A ELISA was 2.5 ng.

PAGE and immunoblotting of surfactant proteins.

Aliquots of BAL were centrifuged at 150 × g for 10 min and the resulting supernatant was centrifuged at 100,000 × g for 45 min. The pellet (purified surfactant) was solubilized, reduced, and fractionated by electrophoresis on a 10% polyacrylamide gel (31). Immunoblotting of separated proteins was done using successive incubations with primary antisera (1:5000) for 1 h and with horseradish peroxidase-conjugated secondary antibodies for 1 h. Bands were visualized by enhanced chemiluminescence.

For analysis of SP-D, the 100,000 g supernatant was run through a 1-mL column of mannose-Sepharose 4B in the presence of 2 mM calcium and captured SP-D eluted with 2 mM EDTA, concentrated on a Centricon 10 concentrator (Amicon, Beverly, MA) and subjected to PAGE and immunoblotting as reported above.

Morphologic Analyses

Immunofluorescence.

Immunochemical staining of immersion-fixed and paraffin- embedded lung biopsy material was performed as previously published (32,33). Fluorescence images of air-dried and Mowiol-mounted slides were viewed on an Olympus I-70 inverted fluorescence microscope with filter packages High Q FITC for EGFP (excitation at 480 nm, emission at 535/550 nm), and High Q TR for Texas Red (excitation at 560/555 nm, emission at 645/675 nm) obtained from Chroma Technology (Brattleboro, VT). Image acquisition, processing, and overlay analysis were performed using IMAGE1 software (Universal Imaging, West Chester, PA).

Immunoelectron microscopy.

For immunoelectron microscopy, lungs were prepared as described previously (34). Briefly, small biopsies were fixed by immersion in a mixture of 4% paraformaldehyde and 0.1% glutaraldehyde in 0.2 M HEPES buffer (pH 7,4). The tissue blocks were infiltrated with 2.3 M sucrose for 3–4 h and frozen in liquid nitrogen. The frozen samples were transferred to 0.5% uranyl acetate in methanol at −90°C for at least 36 h. The temperature was raised to −45°C at a rate of 5°C/h. The samples were washed several times with pure methanol and transferred to Lowicryl HM 20 via HM20-methanol 1:1 and 2:1 for 2 h each. Blocks were polymerized under UV light for 2 d at −45°C. Ultrathin sections were labeled with monospecific surfactant antisera as described (34).

In Vitro Analysis of SP-C Expression

SP-C and EGFP / hSP-C expression constructs.

A full-length human SP-C cDNA insert of 875 bp corresponding to the published sequence (Group II clones) of Warr et al. (26) was the generous gift of Dr. Philip Ballard. This insert was subcloned into the pcDNA3 eukaryotic expression vector polylinker at the EcoRI site as previously described for rat SP-C (35). The resultant wild-type human SP-C clone [termed pcDNA3-hSP-C (6+)] was then used as template for generation of a chimeric fusion protein consisting of EGFP and wild-type human SP-C (EGFP/hSP-C1–197) via a PCR as previously described (19,36).

To generate the mutant EGFP/hSP-CL188Q, a two-primer, single-reaction PCR technique was also used with pcDNA3-hSP-C (6+) serving as template. The primers are shown in Table 1. Amplification produced a purified PCR insert containing the desired point mutation at codon 188 and lacking the 3′ untranslated region, which was ligated into pEGFP-C1 after digestion with BamHI and BglII.

Table 1 Primer sets used in PCR reactions for generation of EGFP/hSP-C fusion constructs
Full size table

For construction of EGFP/hSP-CE66K, mutagenesis was also performed by overlap extension PCR with the primer sets listed in Table 1. The resulting mutant insert, hSP-CE66K containing the 3′ untranslated region was purified and ligated into pEGFP-C1 after digestion with BglII and BamHI.

The sequence fidelity of each generated insert was confirmed by automated DNA sequencing.

A549 cell line and transfection.

The human lung epithelial cell line A549 originally obtained through the American Type Culture Collection (Manassas, VA) has been used in prior studies (19,32,33,35,36). A549 cells grown to 50% confluence on glass coverslips were transiently transfected with EGFP/SP-C constructs (10 μg/ dish) CaPO4 precipitation as previously described (32). Transfection efficiencies ranged from 10 to 20%.

For studies of permeabilized cells, fixation by immersion of coverslips in 4% paraformaldehyde was followed by incubation with primary antisera for 1 h at room temperature followed by either secondary goat anti-mouse IgG monoclonal or secondary goat anti-rabbit IgG polyclonal antisera (32).

RESULTS

Sequence analysis of genomic DNA from the index patient revealed a heterozygous mutation in exon 2 of the SP-C gene (position 1509 G>A) leading to GAG to AAG that results in substitution of a lysine for glutamic acid at codon 66 (SP-CE66K). The remainder SP-C sequence (exons 1–5) for both alleles corresponded to the SP-C sequence submitted by Glasser (GenBank Accession # J03890). The E66K substitution was not present in either parent and was not found on 296 alleles from controls without lung disease indicating it was not a common polymorphism. Both parents were without pulmonary symptoms to date and had no history of previous lung disease.

The expression of SP-CE66K was associated with the presence of interstitial lung disease in the index patient. Hematoxylin and eosin staining of paraffin-embedded sections from an open-lung biopsy obtained at 2 mo of age (Fig. 1, a and b) revealed the presence of cellular NSIP with thickened alveolar septa containing variable amounts of collagenous fibers. The cellular component of the intraalveolar septae consisted mainly of vimentin positive mesenchymal cells (Fig. 1c). Immunostaining for TTF-1, LCA (leukocyte common antigen), and CD3 revealed a hyperplasia of positive type II pneumocytes and the presence of a mild interstitial chronic inflammation with mainly CD3-positive T-lymphocytes (not shown). The interstitial lung disease was persistent and progressive. High-resolution CT scanning performed at 2 y of age revealed the presence of interstitial infiltrates with some ground opacities (Fig. 2).

Figure 2
figure 2

High-resolution CT scanning using intravenous contrast performed on the index patient at 24 mo of age revealed the presence of bilateral interstitial pneumonia.

Full size image

In addition to interstitial pneumonia, biochemical analysis of BAL revealed high levels of protein and phospholipid accompanied by an elevated concentration of SP-A (Table 2). In Figure 3a, thin layer chromatography of phospholipids performed on three BAL samples, collected at 2-wk intervals, showed, with respect to a patient without interstitial lung disease (control) (Fig. 3b), the complete absence of PG and an increase of sphingomyelin. The majority of PC was present in a disaturated form with palmitic acid representing 69.5% of fatty acids associated with this phospholipid. Apart from the lack of PG and lysoPC, the phospholipid composition of our patient was similar to that of a patient with alveolar proteinosis (Fig. 3c). By Western blotting, SP-A, SP-B, mature SP-C, and SP-D were all present in the BAL (Fig. 4a) and mature SP-C in both controls and index patient samples migrated in a similar fashion (Fig. 4b). Repeated surveillance failed to document the presence of Pneumocystis or other opportunistic infections. Conventional bacterial and viral cultures were also negative, demonstrating a lack of infectious organisms in this patient.

Table 2 Composition of BAL fluid
Full size table
Figure 3
figure 3

Surfactant phospholipid analysis. Surfactant (200 μg phospholipid) from the index patient (a), from an infant undergoing BAL to exclude infection (b), and from of an adult patient with alveolar proteinosis (c) were each extracted and the lipids subjected to two-dimensional thin layer chromatography as described in “Methods.” OR, origin; SM, sphingomyelin; PI, phosphatidylinositol; PS, phosphatidylserine; LPC, lysophosphatidylcholine; PE, phosphatidylethanolamine; CL, cardiolipin; PGL, phosphatidylglycerol; X, unknown,

Full size image
Figure 4
figure 4

Western blotting for surfactant proteins. (a) Total protein from BAL from the index patient was separated by SDS-PAGE and transferred to nitrocellulose as detailed in “Methods.” Each lane was probed with monospecific antisera against SP-A, SP-B, mature SP-C, or SP-D and immunoreactive proteins were visualized by chemiluminescence. The position of molecular weight standards is shown at right. The amounts of BAL used per lane were 10 μL for SP-A and SP-B, 150 μL for SP-C, and 240 μL for SP-D. (b) Western blotting of total BAL protein for mature SP-C was performed on the index patient (lane 2) and compared with BAL recovered from three controls performed to exclude infection (lanes 1, 3, 4). Each lane contained 30 μg of total protein. Mature SP-C was detected as a low molecular weight monomer in all four samples (arrowhead).

Full size image

The distribution of the surfactant proteins in the lung of the index patient was assessed by immunofluorescence. Paraffin sections of the index patient (obtained from an open-lung biopsy) were stained for human proSP-C. Two populations of type 2 cells were identified. In approximately 90% of type II cells, proSP-C expression was detected in small endosome-like peripheral vesicles as well as some minimal staining at the level of the plasma membrane (Fig. 5a, arrowhead). In contrast, in tissue sections from an open lung biopsy from a patient with BPD (“BPD,” Fig. 5c) and a specimen from a normal donor lung graft (“Donor,” Fig. 5e), wild-type proSP-C was associated with apical vesicles of type II pneumocytes. In about 10% of type II pneumocytes, a granular staining pattern consisting of cytosolic deposits of proSP-C with a nominal diameter of ~2 μm was observed (Fig. 6A). Rarely, larger sizes (diameter ~7 μm) were detected (not shown). Corresponding to the immunohistochemical staining pattern for proSP-C, abnormal deposits of small vesicles were found in severely damaged type II pneumocytes (10% of total type II cells examined) at the ultrastructural level (Fig. 6B). To further characterize these forms, an immunoelectron microscopic analysis was performed in which strong labeling of 10 nm gold particles for proSP-C was found over these regions (Fig. 6, C and D). By contrast, SP-B (10-nm gold particles) was found in multivesicular bodies (Fig. 6E) and over the projection core of lamellar bodies in type II pneumocytes (Fig. 6F).

Figure 5
figure 5

Expression of mutant SP-CE66K in lung tissue. Fluorescence (a, c, e) and phase contrast (b, d, f) micrographs of paraffin lung sections from the index patient with heterozygous expression of the E66K mutation (a, b), from a recipient lung of a patient with BPD undergoing bilateral lung transplantation (c, d), and from a normal donor lung (e, f) stained for proSP-C using anti-NPROSP-C as described in “Methods.” Type II cells from the E66K mutant lung (a) exhibit proSP-C localization in small subplasma membrane vesicles (inset, arrowhead). In a minority of cells, larger aggregates of proSP-C were also observed (inset, asterisk). In both normal donor lung and BPD lung, proSP-C expression was limited to large cytoplasmic organelles at the apex of the cell (inset, c and e, arrows).

Full size image
Figure 6
figure 6

hSP-CE66K accumulates in nonlamellar body intracellular compartments. (A) Immunostaining of lung section from index patient for proSP-C revealed a granular immunohistochemical pattern in some cells with formation of aggregates of proSP-C (arrow). (B) Ultrastructural analysis using EM on a specimen prepared from the same sample showing the presence of both lamellar bodies and amorphous intracellular aggregates in small vesicles (asterisk) in severely damaged type II pneumocytes. (C–F) Immunogold EM was performed on the same biopsy specimen from the index patient as described in “Methods.” Using anti-NPROSP-C and 10-nm gold particles, a strong labeling for proSP-C was observed over the aggregates (C, asterisk). Nearly all aggregates containing proSP-C were found in close proximity to a swollen endoplasmic reticulum (D, double arrows). Substituting a polyclonal anti-SP-B antibody with 10-nm gold particles (E–F), SP-B isoforms were detected in multivesicular bodies (E, double arrow) and over the projection core of lamellar bodies in type II pneumocytes (F, asterisk).

Full size image

To directly assess the functional significance of the E66K substitution, in vitro trafficking studies were performed utilizing lung epithelial cells transfected with expression constructs encoding for fusion proteins of EGFP in either wild type SP-C or the E66K mutant form (Fig. 7). Control transfections using EGFP/hSP-C1–197 revealed protein expression in cytosolic vesicles of A549 cells. This compartment is CD63 positive (Fig. 7) and has been previously shown to be EEA-1 negative (19), consistent with targeting of wild-type fusion protein to acidic lysosomal-like organelles. Trafficking of the EGFP/hSP-CE66K mutant was markedly different, with distribution mainly in small cytosolic vesicles that were CD63 negative, EEA-1 positive consistent with accumulation in early endosomes (Fig. 7).

Figure 7
figure 7

Expression of EGFP tagged human SP-C mutants in vitro. A549 cells grown on glass coverslips to 50–60% confluence were transfected with 10 μg each of plasmids containing cDNA constructs for expression of EGFP fusion proteins containing either wild-type SP-C (hSP-C1–197) (columns 1, 3), hSP-CE66K (columns 2, 4, 6), or hSP-CL188Q (column 7) using CaPO4 precipitation as described in “Methods.” Forty-eight hours after transfection, cells were fixed, permeabilized, and stained with primary monoclonal EEA-1 antiserum, polyclonal antiCD-63, or monoclonal anti-ubiquitin followed by IgG-specific secondary Texas Red (TR)-labeled antisera. Images were acquired by video fluorescence microscopy using High Q FITC filter package for EGFP fusions or a High Q Texas Red filter package for marker antibody labeling. When subjected to color overlay using IMAGE 1 software (third row), areas of colocalization appear as yellow. Corresponding phase images for each set are shown at the bottom. The majority of EGFP/hSP-CE66K was associated with EEA-1 staining while EGFP/hSP-CL188Q formed ubiquinated perinuclear aggregates. Images are representative of 10 separate experiments for each construct and >50 cells for each experiment.

Full size image

Two previously published SP-C mutations both located in another region of the proSP-C molecule have also been associated with interstitial lung disease (L188Q and ΔExon 4) (16,37). When compared with hSP-CE66K, expression of EGFP/ hSP-CL188Q produced a different subcellular distribution in vitro. In A549 cells, this mutant developed perinuclear aggregation with ubiquitination of the accumulated protein (Fig. 7). In addition, similar to features observed previously with the hSP-CΔExon4 mutant (19). EGFP/ hSP-CL188Q did not colocalize with CD63 or EEA-1 (data not shown). Importantly, EGFP/hSP-CE66K was ubiquitin negative (Fig. 7, right panels) indicating that mutations located in two spatially distinct regions of the proSP-C molecule (proximal versus distal COOH propeptide) yield two distinct trafficking patterns in vitro.

DISCUSSION

Over the past two decades, our understanding of the biosynthetic pathway for SP-C has greatly improved. The proteolytic cleavage steps for processing of the SP-C propeptide, the structural domains that mediate protein trafficking, the subcellular compartments involved, the importance of proper protein folding, and the presence of an intact regulated secretory pathway for successful biosynthesis are now well accepted (11,12,19,33,35,36). It has been demonstrated that alterations in the proSP-C sequence that result in either misfolding or mistargeting can induce formation of intracellular aggregates and/or ER retention. Recently, we have shown that a deletion of the exon 4 transcript of the SP-C gene found in association with idiopathic interstitial pneumonia (17) is characterized by similar defects in intracellular trafficking in vitro (19). The present study extends these observations through the characterization of a structurally distinct mutation in the SP-C gene, which, when detected in biopsy specimens or expressed in vitro, is associated with abnormal targeting of the protein and accumulation in early endosomal compartments. Thus, in contrast to previously described SP-C mutations, the E66K missense substitution (hSP-CE66K) represents a class of spatially and functionally distinct mutations of proSP-C capable of producing chronic lung disease but are phenotypically distinct from their predecessors.

For many diseases, the homozygous expression of a mutant protein induces a global loss of function by interference with protein trafficking or function. In contrast, the heterozygous expression of a mutant allele is sometimes sufficient to induce pathology as represented by a diverse set of chronic, degenerative disorders in a variety of organ systems including brain, liver, and lung. Alzheimer's disease, Parkinson's disease, alpha-1 antitrypsin, amyloidosis, spinocerebellar ataxia, and Charcot-Marie-Tooth disease are all associated with expression of a mutant protein that undergoes a fatal conformational rearrangement (38). At both a recent international conference on α-1 antitrypsin deficiency as well as a workshop on protein processing and degradation in the lung sponsored by the National Heart, Lung, and Blood Institute, the mechanisms by which cells respond to production of abnormal protein conformers and how such misfolded proteins cause cytotoxicity were highlighted as important emerging questions (39,40). Cells have a limited repertoire for disposal of abnormal products. Normally, degradation of both integral membrane proteins as well as secreted peptides unable to fold in the ER or to traffic anterograde involves one of four pathways: 1) retrograde routing to cytosolic proteasomes, 2) diversion to the constitutive secretory pathway, 3) accumulation in endosomes/lysosomes, or 4) development of autophagosomes. Data showing a tendency of mutant proSP-C to either aggregate or be rerouted to any of these compartments would be consistent with this concept of conformational disease.

The role of SP-C mutations in producing conformational mediated disease in the lung was initially supported by two early reports of patients with heterozygous expression of TWO different mutant proSP-C forms in association with interstitial lung disease (16,17). The first SP-C mutation, c460 + 1 G>A, was found on one allele both in a full-term infant with histologic features of NSIP and in the mother whose lung pathology was consistent with UIP produced alternate splicing of the SP-C mRNA, deletion of exon 4, and resultant production of a defective proprotein foreshortened by 37 amino acids (17). The second mutation was an exon 5 T>A transversion, which substitutes glutamine for a conserved leucine at residue 188 of the COOH propeptide (“L188Q”), was present in a kindred with histologic features of UIP in adults and NSIP in children (16). Of note, the distal COOH region of proSP-C (F94-I197) has recently been structurally linked with an emerging family of several previously unrelated proteins all containing a novel BRI domain (BRICHOS) (18). The exact function of BRI family members is unknown, however, degenerative dementia with CNS amyloid formation in kindreds expressing BRI mutations is well described (41). Using EGFP fusion proteins, we have demonstrated induction of ubiquinated protein aggregates in vitro by both hSP-CΔExon4 (19) and hSP-CL188Q (Fig. 7), which have subsequently been shown to be toxic to the developing mouse lung in vivo (20).

In contrast to these mutations, the E66K substitution in proSP-C lies more proximal to BRICHOS and as such would not be expected to disrupt cysteine-mediated folding, as has been the case for either hSP-CL188Q or hSP-CΔExon4. Expression of hSP-CE66K in vitro was not associated with the formation of ubiquitin-labeled aggregates but instead was found in EEA1-positive vesicles. Furthermore, the clinical picture of the patient differed from the previous patient series because the interstitial pneumonia was complicated by coexistence of alveolar lipoproteinosis and the presence of intracellular proSP-C aggregates (Fig. 5). Interestingly, it has been recognized that patients with Hermansky-Pudlak syndrome (HPS) can develop pulmonary fibrosis (42). HPS is a polygenetic disorder associated with defective vesicle formation including abnormal melanosomes, platelet granules, lysosomes, and lamellar bodies (43). Mice with HPS1 and HPS2 gene mutations demonstrate increased lung levels of lysosomal enzymes and phospholipids in conjunction with fibrosis suggesting that altered vesicular trafficking and accumulation of abnormal protein can promote lung inflammation and fibrosis (44).

In the current work, the BAL of the index patient contained detectable mature SP-C. This is in contrast to patients with heterozygous expression of the hSP-CΔExon4 allele in which mature SP-C was absent suggesting the presence of a dominant negative effect upon SP-C biosynthesis by that mutant allele (17). Because SP-B was present in the patients with the Δexon4 mutation, the contribution of the lack of SP-C to the underlying lung pathology is unclear. The SP-C null mouse has been shown to be viable (14), however, a recent report suggests that when SP-C −/− mice are back-crossed onto a different strain, they develop emphysema (15). Because SP-C was still detectable in the current patient, it is unlikely that the pathology of interstitial pneumonias (UIP, NSIP) associated with SP-C mutations is due simply to a lack of SP-C (dominant negative effect). More plausible, the data suggest that SP-C mutations induce a toxic gain of function that is a consequence of the resultant abnormalities in protein trafficking.

The E66K substitution in the SP-C gene of this child is the result of a sporadic mutation. The clinical picture of hSP-CE66K is similar to that of two other recently reported human SP-C mutations (45,46). A sporadic substitution of threonine for isoleucine at codon 73 was noted in an infant with well-characterized NSIP pattern on histopathology and PAS (periodic acid-Schiff)-positive material (45). In vitro expression of hSP-CI73T/ EGFP fusion protein in A549 cells was associated with deposition in early endosomes. A recent preliminary study has reported additional heterozygous hSP-CI73T mutations in association with interstitial lung disease, some of which occurred spontaneously in several unrelated patients (37). In addition to multiple cases of I73T substitutions, an in-frame 9-bp deletion spanning codons 91–93 in exon 3 of the SP-C gene present on one allele was described in a patient with ILD undergoing lung transplantation. Neither parent carried this deletion. By immunofluorescence, proSP-C in the transplanted lung was aggregated within type II cells in a compartment separate from SP-B (46). Taken together, the de novo occurrence of all three of these mutations (hSP-CE66K, hSP-CI73T, and hSP-CΔ91–93) in relation to the development of chronic lung disease and located within 30 amino acids of each other, lends support to the pathogenicity of mutations in the proximal COOH flanking region. Furthermore, the nonBRICHOS portion of proSP-C (H59-T93) could represent a genetic hot-spot for spontaneous mutations in the SP-C gene.

Analogous to what has been developed for understanding the pathophysiology of the cystic fibrosis transmembrane receptor, a working model to explain the contrasts between phenotypes found with the two major forms of SP-C mutations reported to date is shown in Figure 8. In normal biosynthesis, proSP-C is trafficked as a homodimer from the ER through the Golgi to the multivesicular body and ultimately to the lamellar body where the mature protein is secreted with phospholipid and SP-B. In contrast, expression of mutant forms of SP-C can result in one of two fates. In the first pathway, BRICHOS mutants (such as hSP-CΔ188Q) are recognized by ER quality control mechanisms. Termed ER-associated degradation (ERAD) this pathway involves the ubiquitination of the abnormal protein and targeting to the proteasome for disposal. If overwhelming to the proteasome, aggregated proteins can be diverted retrograde to a microtubule-dependent inclusion body termed the aggresome (47). Because of the oligomerization, the wild-type allele product is also routed here leading to a dominant negative effect imposed by the ubiquitinated mutant resulting in a lack of SP-C (19). However, in addition, there appears to be a toxic gain of function. Retention of protein in the ER can itself initiate an inflammatory response (ER overload response) (48). Recent data by Bridges et al. (20) suggests that overexpression of hSP-CΔΕxon4 in vitro can activate a Bip-dependent ER overload response.

Figure 8
figure 8

Model for aberrant trafficking of human SP-C mutants associated with ILD. Schematic diagram showing possible fates for expressed human SP-C proteins. Normal SP-C biosynthesis proceeds via sorting and delivery of proSP-C21 to cytosolic processing sites in the regulated secretory pathway including ER, Golgi, cytosolic vesicles, multivesicular bodies (MVB), composite bodies (CB), and lamellar bodies (LB) where it is co-secreted with SP-B and lipid (solid arrows). Reuptake of phospholipid and surfactant proteins occurs via endocytosis, early endosome incorporation, and recycling via MVB or degradation in lysosomes. The previously published mutants hSP-CL188Q and hSP-CΔExon4 overwhelm ER quality control mechanisms forming large toxic intracellular aggregates (aggresomes) (dotted arrows). In contrast, mutants of the proximal flanking COOH propeptide such as hSP-CE66K are diverted from normal biosynthesis and accumulate in the endosomal pathway directly or after deposition in the plasma membrane (dashed arrows) and inhibit the recycling pathway. UPR, unfolded protein response; ibm180, ABCA3 gene produt; LAMP-1, lysosome associated membrane protein.

Full size image

In contrast to BRICHOS mutations, for E66K as well as other mutants within the proximal COOH domain, aggresome formation does not occur. We propose a second pathway, whereby substitutions in the nonBRICHOS region (e.g. hSP-CE66K) result in a milder protein phenotype not severe enough to grossly misfold and trigger the ER stress response (Fig. 8). Because proSP-C is processed as an integral membrane protein, diversion of the protein away from the regulated secretory pathway would limit the protein's ability to be properly processed. Thus, by either fusion of the constitutive vesicles containing membrane bound mutant proSP-C to the plasma membrane and bulk membrane recycling mechanisms or transport of mutant proSP-C directly from the trans-Golgi, endosomal accumulation can occur (49). Unlike either normal anterograde trafficking or ubiquitination-mediated proteasomal routing, this entry into default pathways does not require a strong trafficking signal. Heteromers containing wild-type protein would present an overriding signal to continue anterograde, and mutant protein is unlikely to function as a dominant negative. We have shown in vitro that a different experimental SP-C mutant that lacks NH2 targeting signals can be trafficked anterograde via heteromeric association with co-transfected wild-type protein (36). Thus, in parallel with abnormal diversion of mutant hSP-CE66K to endosomes, a portion of both the mutant and wild-type isoforms could also undergo normal trafficking and processing as heterodimers to the mature protein and SP-C3.7 has been detected in lavage. Importantly our findings regarding hSP-CE66K indicate that the pathogenicity from expression of this mutant cannot come from the lack of SP-C but rather represent a disruption of normal cellular function by accumulation of mutant forms in atypical intracellular compartments (toxic gain of function).

Our immunocytochemistry indicates that aggregation is occurring in the histopathology sections. We have noted that up to 10% of the cells contain intracellular aggregates of proSP-C. There are two possibilities for the heterogeneity in the cellular expression pattern. Different alleles are somatically expressed in different type II cells. Alternatively, temporal and spatial heterogeneity in the level of expression of both alleles is occurring. Given that the biopsy specimen represents but a single time point, the dynamic formation of intracellular aggregates is not experimentally assessable from this data. The transmission of proteotoxicity across cellular compartments is an emerging area of investigation. Recently, it has been shown by Jin et al. (50) that intracellular accumulation of amyloid precursor protein disrupts endosome function in the Niemann-pick type C disease. Given the similar amyloidgenic properties of SP-C (51–53), it seems plausible that endosomal accumulation and aggregation of proSP-C leading to epithelial cell dysfunction results via a similar pathway. The accumulation of mutant protein in early endosomes as we observed could result in impairment of endocytosis leading to the extracellular accumulation of phospholipids and SP-A found in this patient.

In summary, we have characterized the phenotype of a patient with interstitial pneumonia, alveolar proteinosis, and a heterozygous mutation in the surfactant protein C gene. In contrast to previously reported SP-C mutations, this substitution represents a novel functional isoform in a spatially distinct region of the molecule. Both in vitro and in vivo data are compatible with diversion of the mutant protein to early endosomes and other nonlamellar body compartments. Further details of the molecular mechanisms translating the expression of the mutant protein in this compartment with the development of pulmonary fibrosis and lipoproteinosis will require development of long-term expression models such as transgenic mice, which are currently underway.