The common ABCA3^E292V variant disrupts AT2 cell quality control and increases susceptibility to lung injury and aberrant remodeling

Abstract

ATP-binding cassette class A3 (ABCA3) is a lipid transporter that plays a critical role in pulmonary surfactant function. The substitution of valine for glutamic acid at codon 292 (E292V) produces a hypomorphic variant that accounts for a significant portion of ABCA3 mutations associated with lung disorders spanning from neonatal respiratory distress syndrome and childhood interstitial lung disease to diffuse parenchymal lung disease (DPLD) in adults including pulmonary fibrosis. The mechanisms by which this and similar ABCA3 mutations disrupt alveolar type 2 (AT2) cell homeostasis and cause DPLD are largely unclear. The present study, informed by a patient homozygous for the E292V variant, used an in vitro and a preclinical murine model to evaluate the mechanisms by which E292V expression promotes aberrant lung injury and parenchymal remodeling. Cell lines stably expressing enhanced green fluorescent protein (EGFP)-tagged ABCA3 isoforms show a functional deficiency of the ABCA3^E292V variant as a lipid transporter. AT2 cells isolated from mice constitutively homozygous for ABCA3^E292V demonstrate the presence of small electron-dense lamellar bodies, time-dependent alterations in macroautophagy, and induction of apoptosis. These changes in AT2 cell homeostasis are accompanied by a spontaneous lung phenotype consisting of both age-dependent inflammation and fibrillary collagen deposition in alveolar septa. Older ABCA3^E292V mice exhibit increased vulnerability to exogenous lung injury by bleomycin. Collectively, these findings support the hypothesis that the ABCA3^E292V variant is a susceptibility factor for lung injury through effects on surfactant deficiency and impaired AT2 cell autophagy.

INTRODUCTION

The lipid transporter adenosine triphosphate (ATP)-binding cassette subfamily A, member 3 (ABCA3) glycoprotein has been increasingly recognized as an essential component for the proper biogenesis of pulmonary surfactant. Located in the limiting membrane of intracellular vesicles of alveolar type 2 (AT2) cells including lamellar bodies (LBs), ABCA3 functions as a lipid pump (1–6). Various in vitro and in vivo studies have shown that cholesterol and phospholipids including phosphatidylcholine, phosphatidylglycerol, phosphatidylserine, sphingomyelin, and phosphatidylethanolamine are among the substrates of ABCA3 (1–4). As seen in Abca3 knockout mouse model studies and human ABCA3-null patients who die from surfactant deficiency/respiratory distress soon after birth, the transporter is recognized as a critical regulator of LB biogenesis and surfactant function (1, 3, 4, 7).

Beyond lethal/null phenotypes, a growing body of evidence indicates that some mutations in the ABCA3 gene are associated with chronic postneonatal lung disease including childhood interstitial lung disease (chILD), as well as diffuse parenchymal lung disease (DPLD) and idiopathic pulmonary fibrosis (IPF) in adults. In total, more than 200 sporadic and familial mutations consisting of homozygous, missense, nonsense, and frameshift mutations, as well as heterozygous insertion and splice-site mutations, have been reported to date in association with both null and chILD/DPLD/IPF phenotypes. According to the Genome Aggregation Database (gnomAD) (https://gnomad.broadinstitute.org/gene/ENSG00000167972), over a thousand missense variants in ABCA3 have been identified to date. It has also become increasingly evident that additional cis- and trans-allelic mutations within the ABCA3 gene may act as disease modifiers (8–11). Many of these patients suffer from chronic lung disorders (with diagnoses of various types of DPLD including pulmonary fibrosis) and carry compound heterozygous ABCA3 mutations [for review, see Ref. (12)]. These findings suggest that although the presence of a hetreozygous ABCA3 mutation may not cause a readily recognizable phenotype or lead to mild lung abnormalities that escape detection, homozygosity, compound heterozygosity for ABCA3 mutations, or other modifier genes in addition to environmental factors may influence the severity of the disease (12).

The valine for glutamic acid missense substitution at residue 292 (E292V) in the highly conserved first cytosolic loop is the most common ABCA3 mutation associated with various lung abnormalities in children and adults (7, 13–18). It is present in ∼0.3%–0.4% of the general US population of European descent (including healthy individuals and those with lung disease) (15) and 1.3% of individuals in a study conducted in the general Danish population (19). The Danish study ruled out heterozygosity of the ABCA3 E292V mutation as a major risk factor for reduced lung function or for an increased risk of chronic obstructive pulmonary disease (COPD) in this population. In vitro studies have shown that the E292V variant is a functional hypomorph exhibiting moderately depressed ATP-binding and lipid transport activities (17). This is also the most common variant among individuals harboring compound heterozygous ABCA3 mutations associated with lung disease (12). Homozygosity for the E292V variant in these populations shows variable penetrance with clinical presentations ranging from lung dysfunction to diffuse parenchymal disorder during adulthood (18). In some patients, severe outcomes are evidenced by death from respiratory failure shortly after birth (18) or the requirement for lung transplantation due to severe lung dysfunction. Collectively, findings thus far suggest that the E292V mutation, particularly the expression of the heterozygous variant by itself, poses a relatively low risk to lung health and that carriers are commonly asymptomatic. Given the frequency of the E292V mutation in the general population, especially as it relates to lung disorders, we sought to characterize the mechanisms by which ABCA3 mutations such as E292V alter basic biological homeostasis of AT2 cells/lung to identify the key mechanisms triggered that could contribute to the etiology of lung pathology seen in these patients.

Informed by a case in a young adult patient homozygous for the ABCA3 E292V mutation, we used both in vitro modeling and in vivo studies in an Abca3 knock-in murine model constitutively expressing the mutant isoform. Using cell lines transiently or stably expressing enhanced green fluorescent protein (EGFP)-tagged ABCA3 constructs, we extended the prior observation of a functional defect in the mutant ABCA3 transporter (17). Phenotyping of mice homozygous for E292V (E292V^+/+) revealed multiple abnormalities including age-dependent alteration of macroautophagy (hereafter referred to as autophagy) accompanied by augmented mitochondrial mass and increased caspase 3 activation in AT2 cells, as well as age-dependent alveolitis along with elevated alveolar collagen deposition and a vulnerability to exogenous lung injury. Our results further support the role of mutation-induced altered AT2 cell homeostasis as a susceptibility factor for aberrant lung injury/repair/remodeling.

Clinical Presentation of a Patient Homozygous for the ABCA3^E292V Mutation

A male of European descent without a history of cardiopulmonary disease presented at age 25 yr for evaluation of dyspnea. His family history was negative for pulmonary disease, and he did not use tobacco products. He was found on pulmonary function testing to have restrictive lung physiology with a diffusion impairment, and subsequent chest radiography revealed an upper-lobe-predominant diffuse parenchymal lung disease (DPLD), which progressed over time and was characterized by peribronchial thickening, scattered ground glass opacities, and peripheral reticulations with traction bronchiectasis and bronchiolectasis (Fig. 1, A and B). With progressive respiratory failure, he underwent bilateral lung transplantation at age 47 yr. His explanted lung pathology revealed peripheral dilated airspaces with septal thickening and distortion from smooth muscle expansion into the alveolar septa (Fig. 1C). Further pathological examination revealed spatially heterogeneous lung remodeling with regions of distal lung bronchiolarization, hyperplasia of the AT2 cell, and accumulation of proteinaceous material and foamy macrophages in the distal airspaces (Supplemental Fig. S1; all supplemental material is available at https://doi.org/10.6084/m9.figshare.14527098).

Figure 1.

Pretransplant radiography and explant histology from a 47-yr-old patient with DPLD harboring the homozygous E292V variant. A: chest radiograph before transplant showing a diffuse reticular pattern throughout the bilateral lungs. B: chest CT before transplant demonstrating DPLD characterized by architectural distortion with peribronchial thickening with consolidation, diffuse reticular changes, and traction bronchiectasis and bronchiolectasis. C: lung explant histology: H&E staining at low (left) and high (right) magnifications showing end-stage lung disease with diffuse parenchymal remodeling characterized by dilated airspaces with thickened septa replacing the normal distal lung alveolar architecture (left) and extension of smooth muscle into the alveolar septa (right). CT, computed tomography; DLPD, diffuse parenchymal lung disease; H&E, hematoxylin-eosin.

MATERIALS AND METHODS

Materials

The pEGFP-N1 monomer plasmid was purchased from Clontech, Inc. (Palo Alto, CA). Tagged constructs were necessary to follow trafficking of the ABCA3 transporter since we were unable to find reliable antibodies against the protein. Tissue culture medium was produced by the Cell Center Facility, University of Pennsylvania. Except where noted, all other reagents were electrophoresis, tissue culture, or analytical grade and were purchased from Sigma Chemical, Inc. (St. Louis, MO), or BioRad, Inc. (Melville, NY).

Generation of Mouse pgk-Neo-ABCA3-E292V Mutant Targeting Vector

Construction of the targeting vector was performed by Gene Bridges Company (Heidelberg, Germany) using a recombineering-based strategy as previously reported (20, 21).

Production of the ABCA3^E292V Mouse Line

Using this vector, ABCA3^E292V mouse lines were generated in collaboration with the University of Pennsylvania Gene Targeting Core and Transgenic and Chimeric Mouse Facility. Briefly, the targeting vector was linearized by using unique NotI and SacII restriction sites and electroporated into C57BL/6 mouse embryonic stem (ES) cells. Following G418 selection, surviving colonies were picked and screened for homologous recombination using a primer specific for the targeted allele. The presence of the targeted allele was confirmed by Southern blot hybridization in 11 clones. Two targeted ES cell clones containing the pgk-Neo cassette together with the E292V mutation were injected into blastocysts of BALB/c mice and transferred to pseudopregnant females. Chimeric offspring were crossed with C57BL/6 mice, and their offspring were screened for germline transmission by coat color, PCR, and Southern blot hybridization. As shown in Fig. 3A, the F1 generation containing the pgk-Neo cassette (inserted in reverse orientations, rNeo) was crossbred with transgenic C57BL/6J mice constitutively expressing an enhanced variant of Saccharomyces cerevisiae FLP1 recombinase (FlpE) (Jackson Laboratory, Bar Harbor, ME) to remove the pgk-Neo cassette and to produce the ABCA3^E292V mouse line.

All protocols were reviewed and approved by the Institutional Animal Care and Use Committees of the University of Pennsylvania and adhered to the principles of the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

ABCA3/EGFP Constructs

The human full-length ABCA3^WT/EGFP was generated by amplifying three overlapping ∼2-kb segments of cDNA by a reverse transcription (RT)-PCR method described previously (22). The plasmid served as a template to generate EGFP fusion E292V mutant construct via the one-step PCR amplification and subcloning method using the QuikChange kit as described earlier.

Cell Lines and Transfection

Human A549 and HEK293 epithelial cell lines were originally obtained through the American Type Culture Collection (ATTC, Manassas, VA). A549 cells grown to 85%–90% confluency on glass coverslips in 35-mm plastic dishes were transiently transfected with fusion wild-type or mutant ABCA3 constructs (4 µg/dish) using Lipofectamine 2000 (Invitrogen, Carlsbad, CA). ABCA3^WT/EGFP and ABCA3^E292V/EGFP stable cell lines in HEK293 cells were generated following a 48-h transient expression using the CaPO₄ transfection method (5) and with subsequent steps taken for clonal selection with 1 mg/mL G418. The cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) containing 500 µg/mL G418, 10% FBS, and 1% penicillin/streptomycin.

Immunostaining

Colocalization studies were performed by immunostaining plated cells or lung sections that were fixed by immersion of coverslips in 4% paraformaldehyde. Following permeabilization, cells were immunolabeled with primary antibodies for 1 h at room temperature at the following dilutions: anti-CD63 (Immunotech), 1:500; anti-calnexin (Stressgen), 1:200; and lysosomal-associated membrane protein 1 (LAMP1) (Abcam), 1:1,000. Alexa Fluor-conjugated secondary goat anti-mouse IgG monoclonal or secondary goat anti-rabbit IgG polyclonal antibodies (Jackson ImmunoResearch Laboratories) at 1:200 dilutions were used for visualization.

Immunofluorescence Microscopy

Epifluorescent images of mouse lung sections immunostained with Alexa Fluor-conjugated anti-proSP-C (NPRO) (in house) were examined as described previously (23). Confocal fluorescent images of permeabilized A549 or AT2 cells (isolated from human fetal lung tissue) fixed in 4% paraformaldehyde and immunostained with Alexa Fluor-conjugated antibodies were examined by a Bio-Rad Radiance 2000 confocal scanning system (Carl Zeiss).

Picrosirius Red Staining

Staining for collagen was performed using the Picrosirius Red Stain Kit (Polysciences Inc.) following the manufacturer’s instructions. Images were viewed on an Olympus I-70 inverted fluorescence microscope using a High Q red fluorescence filter (excitation 560/555 nm; emission 645/675 nm) as described previously (23). Image acquisition was performed using the Metamorph 7.5 software (Molecular Devices, Inc., Downingtown, PA). Digital morphometric measurements of the acquired images were performed by ImageJ using 10 random fields of peripheral and central lung (without large airways) for each section and analyzed at a final magnification of ×400. The mean area of lung field in each section staining for Picrosirius red was calculated and expressed as a percentage of total section area (24, 25).

Immunoblotting

Mouse AT2 cells freshly isolated or following overnight reagent treatment in 35-mm culture dish were counted using an automated cell counter (NucleoCounter, New Brunswick Scientific) and solubilized in sample loading buffer (93.75 mM Tris-HCl at pH 6.8, 3% SDS, 15% glycerol, and 0.015% bromophenol blue) supplemented with 5% dithiothreitol. Samples were sonicated on ice in 10-s bursts twice at 50 W and heated to 95°C for 5 min, and proteins were separated by SDS/PAGE as described previously (26) and transferred to PVDF membranes. Immunoblotting of transferred samples was performed using incubations with primary anti-GFP (Molecular probes), 1:1,500; anti-LC3B (Cell Signaling), 1:1,000; anti-p62/sequestosome-1 (SQSTM1) (Cell Signaling), 1:1,000; anti-translocase of outer membrane 20 (Tom20) (Santa Cruz), 1:1,000; anti-Rab 7 (Cell Signaling), 1:1,000; anti-LAMP1 (Abcam) 1:1,000; anti-full/cleaved caspase 3 (Cell Signaling), 1:1,000; anti-GRP-78 (Abcam) 1:1,000; anti-X-box binding protein 1 (XBP1) (Santa Cruz), 1:200; anti-Hdj-2 (Lab Vision), 1:400; anti-SP-B (PT3) (Ross Laboratories), 1:3,000; anti-SP-C (NPRO) (in house), 1:3,000; anti-ubiquitin (Cell Signaling), 1:3,000; or anti-β-actin (Sigma), 1:5,000; and horseradish peroxidase-conjugated secondary antibodies as previously described (23, 68, 78). Bands visualized by enhanced chemiluminescence (ECL2 No. 80196 Thermo Fisher or WesternSure No. 926-95000, LI-COR Biotechnology, Lincoln, NE) were acquired by exposure to film or direct scanning using a Li-Cor Odyssey Fc Imaging Station and quantitated using the manufacturer’s software.

Bronchoalveolar Lavage Fluid Analysis

Bronchoalveolar lavage fluid (BALF) was collected from mice by gently lavaging five times with 1 mL of sterile saline. Processing and analysis of BALF has been described previously (25). Cytokine levels were determined from cell-free supernate by enzyme-linked immunosorbent assay (ELISA) using antibodies according to manufacturer’s instruction (R&D Systems). Total phospholipid content was determined by Bartlett’s colorimetric estimation of inorganic phosphorus (27). Immunoblotting of BALF was performed according to the previously described procedure (28) using successive incubations with primary polyclonal anti-SP-B (PT3) (Ross Laboratories), 1:3,000, and anti-mature polyclonal SP-C (Seven Hills Bioreagents), 1:2,500, followed by goat anti-rabbit horseradish peroxidase-conjugated secondary antibody, 1:10,000.

Human Fetal Lung AT2 Cell Culture and Transfection

Isolation and culture of human fetal alveolar type II cells were performed as described previously (29). Briefly, enriched populations of epithelial cells were isolated from second-trimester (13–20 wk of gestation) human fetal lung tissue obtained under a protocol approved by the Children’s Hospital of Philadelphia Institutional Review Board. After overnight culture as explants without hormones, the tissue was digested with trypsin, collagenase, and DNAase, and fibroblasts were removed by differential adherence. Nonadherent cells were first transfected with various wild-type or mutant ABCA3 fusion constructs using electroporation with the Amaxa transfection system before plating on glass coverslips in 35-mm plastic culture dishes in Weymouth’s medium containing 10% fetal calf serum. After overnight culture (day 1), attached cells were cultured for an additional 3–9 days in 1 mL of serum-free Weymouth’s medium containing 10 nM dexamethasone, 0.1 mM 8-bromo-cAMP, and 0.1 mM isobutylmethylxanthine (referred to as DCI), previously shown to maximally induce surfactant components consistent with a morphological AT2 cell phenotype.

Mouse AT2 Cell Isolation

Mouse AT2 cells were isolated as previously reported (30) with some modification. Briefly, a single-cell suspension was obtained by instilling dispase (BD Biosciences) into perfused lungs, followed by mechanical dissociation with a McIlwain tissue chopper (Metrohm) and treatment with 20 μg/mL DNAase I (Sigma-Aldrich). Differential adherence on plastic culture dishes negatively selected mesenchymal cells. CD45⁺ cells were depleted by negative selection using Dynabeads Untouched mouse T cells kit (11413 D) and Dynabeads mouse DC enrichment kit (11429 D, Thermo Fisher Scientific). Recovered cells were collected and flash-frozen at −80°C or immediately cultured for experimental procedures. Purity was determined by immunostaining preparations adhered overnight to 10% Matrigel-coated coverslips using DAPI in combination with primary antisera for proSP-C. Purity was assessed by manual counts in five ×20 fields per sample defined as the number of pro-SP-C⁺ cells (AT2 cells) divided by the total nuclei and was routinely >95% (n = 5).

Reagent Treatment of AT2 Cells

AT2 cells freshly isolated from mouse lung cultured in 35-mm plastic dishes were treated with bafilomycin A1 (Baf. A1) (Sigma-Aldrich) and/or rapamycin (Rapa) (LC Laboratories) at indicated concentrations for 18 h.

Sucrose Gradient Fractionation

Lamellar bodies (LB) or lamellar body-like (LB-like) organelles were fractionated from homogenates of either mouse lungs cleared of blood or HEK cells stably expressing the ABCA3^WT or ABCA3^E292V proteins using a sucrose density gradient, as described previously (31).

Quantitative Real-Time Polymerase Chain Reaction

Total lung RNA was purified with TRIzol Reagent (Thermo Fisher Scientific). cDNA was prepared from 500 ng of total RNA using Versa cDNA synthesis kit (Thermo Fisher Scientific). Quantitative real-time polymerase chain reaction (qPCR) was carried out using Fast SYBR Green Master Mix (Thermo Fisher Scientific) in triplicate using mRNA-specific primers and normalized to 18S expression levels using the delta-delta computed tomography (CT) methodology (32). TaqMan assays and primers (Applied Biosystems) for ABCA3 exons 25–27 (Cat. No. 4331182) and exons 11–13 (Cat. No. 4351372) were used. For collagen-related gene primers, Col1A1 (No. Mm00801666_g1, Thermo Fisher) and Col3a1 (No. Mm01254476_m1, Thermo Fisher) were used. Results are depicted as relative quantities of RNA after correcting for 18S followed by normalizing to wild-type controls.

Transmission Electron Microscopy

Transmission electron microscopy (TEM) of HEK cells was performed in the Imaging Core Facility at the University of Pennsylvania School of Medicine using the method of Hayat (33) and modified as published previously (34).

Lung Preparation for Stereological Examination

At the age of 32 wk, whole lungs were fixed by tracheal instillation with a constant pressure of 25 cm H₂O, applying a 1.5% glutaraldehyde/1.5% paraformaldehyde mixture in 0.15 M HEPES buffer and further processed for design-based stereology (35). After determination of lung volumes [V(lung)], complete lungs were subjected to a systematic uniform random sampling procedure for light and TEM as described elsewhere (36), and sampled tissue was embedded in Technovit 8100 (Heareus Kulzer, Wehrheim, Germany) or Epon, respectively. For light microscopy, single sections were stained with toluidine blue. Ultrathin sections of 60 nm thickness of Epon-embedded tissue were cut and placed on one grid for analysis of the lamellar bodies at the electron microscopic level.

Stereological Analysis

Stereological assessment is based on the ATS/ERS statement on quantitative evaluation of lung structure (37, 38). To quantify phenotypic differences at the light microscopic level, surface area of alveolar epithelium [S(alv,lung)], volumes of alveolar septal wall [V(sep,lung)], mean septal wall thickness [τ(sep)], and acinar airspaces of the alveoli [V(alv,lung)] were determined by point and intersection counting (35, 39). Within the reference space, the volume fractions and surface area density of the structures of interest were quantified and multiplied with the reference space (e.g., lung volume) to obtain absolute data per lung.

At the electron microscopic level, LBs were investigated by the point-sampled intercept method from which the volume-weighted mean volume of LBs was determined [ν_V(lb)] (40). In addition, point and intersection counting within the AT2 cells was used to determine the volume fraction and the surface area density of the limiting membrane of the LBs within AT2 cells to calculate the volume-to-surface ratio of LB as a parameter related to the LB diameter.

Bleomycin Lung Injury Model

Mice (12–16, 30–34, 40–44 wk old) were anesthetized with intraperitoneal injection of pentobarbital sodium (70 mg/kg) and suspended by their front teeth from an elastic band attached on an angled plexiglass stand. The tongue was gently lifted with forceps and an otoscope was inserted toward the larynx to allow an unobstructed view of the vocal folds. Then, 50 µL of bleomycin (0.5 or 0.75 U/kg as indicated in figure legend) solution was administered via an insulin syringe fitted with PE-10 tubing, followed by 100 µL of air (Fig. 6). The mice were observed following the procedure to ensure that they recovered completely from anesthesia. At designated timepoints following bleomycin administration, mice were euthanized by pentobarbital injection. Lungs were harvested and prepared for histological examination.

Figure 6.

A: weight loss in E292V^+/+ and WT control mice following intratracheal (IT) bleomycin (0.5 U/kg) instillation. *P < 0.05 vs. WT using unpaired 2-tailed t test. WT, n = 8; E292V^+/+, n = 7. B, top: representative baseline H&E (rows 1 and 3) and trichrome (rows 2 and 4) staining of histological lung sections from 32-wk-old WT and E292V^+/+ mice. Bar, 4 mm. Bottom: fold change (compared with WT average) in collagen genes messages from lung tissues measured by RT-qPCR. *P < 0.01 vs. WT by unpaired 2-tailed t test. WT, n = 5; E292V^+/+, n = 8. C: representative H&E (left column) and trichrome (right three columns) staining of histological lung sections from 32-wk-old E292V^+/+ mice and WT littermates 21 days after IT bleomycin. Patchy trichrome-stained regions were noted in WT mice. More prominent trichrome staining was observed in E292V^+/+ mice displaying severe alveolar destruction and subpleural fibrosis enveloping several lobes. Boxed and numbered regions (left column) are enlarged on the right to provide enhanced resolution. Bar, 3 mm. D: representative distribution of picrosirius red-stained lung section from 32-wk-old WT and E292V^+/+ mice (left) and dot plots of % area of alveolar collagen from transferred images to ImageJ Data software (right), with each dot representing percentage of at least 10 randomly selected fields per mouse lung section (right). *P < 0.05 vs. WT using unpaired 2-tailed t test. n = 6 per group. E, top: representative epifluorescence images of lung sections from 32-wk-old E292V^+/+ mice and WT littermates (untreated or 21 days after IT bleomycin) immunostained for proSP-C. Nuclei are stained with DAPI (blue). Bar, 50 µm; Inset bar, 10 µm. Bottom: dot-plots of average cell number of SP-C⁺ cells per mouse lung obtained from at least 10 fields (each at ×20 magnification field). Alveolar regions absent of large airways were randomly selected for cell count using ImageJ, version 1.8.0, multipoint tool. *P < 0.01 vs. WT Bleo, **P < 0.001 vs. untreated E292V^+/+ by one-way ANOVA with post hoc Tukey’s test. n = 5 per group. F: cell counts (top left) and pulmonary function tests that included static compliance (top right), inspiratory capacity (bottom left), and elastance (bottom right) of E292V^+/+ mice and WT littermates 21 days after IT bleomycin. *P < 0.05 vs. WT using unpaired 2-tailed t test. n = 8 per group. G: Kaplan–Meier survival curve of 42-wk-old E292V^+/+ mice and WT littermates following IT bleomycin treatment (0.75 U/kg). n ≥ 12. End points were defined as death or body weight <75% of starting weight on 2 consecutive days. *P < 0.05 vs. WT by log-rank (Mantel–Cox) test. WT, n = 6; E292V^+/+, n = 8. Bleo, bleomycin; H&E, hematoxylin-eosin; SP-C, surfactant protein; WT, wild type.

TUNEL Assay

Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay was performed using TUNEL assay kit (ab66108, Abcam, Cambridge, UK). Lung sections were fixed for 15 min using 4% paraformaldehyde. Following PBS rinse, the cells were treated with 70% ethanol and incubated for 30 min on ice. After incubation, the cells were washed with buffer three times (1 mL each), and 50 µL of DNA labeling solution was added per lung section, followed by incubation for 1 h at 37°C in a humidified chamber. Following washing steps using rinse buffer three times (1 mL each), immunostaining was performed using NPRO antibody.

Statistics

All data are presented as group means ± SD. Statistical analyses were performed with GraphPad InStat. Student’s t test (unpaired 2-tailed) was used for data describing two groups. Multiple group comparisons were performed by one-way analysis of variance (ANOVA) with the Tukey–Kramer post hoc test. P values < 0.05 were considered significant.

RESULTS

In Vitro Expressed ABCA3^E292V Variant Is Trafficked Normally but Is Functionally Impaired

Published reports of in vitro models of ABCA3^E292V show normal trafficking (17, 41–43) but decreased ATP-binding and lipid transport activities (17, 42). Transient expression of the ABCA3^E292V variant showed normal posttranslational trafficking, localizing to CD63-positive lysosomal organelles in A549 cells and LAMP1 containing lysosomes/LBs AT2 cells isolated from human fetal lung tissue (Fig. 2A). Ultrastructurally, although HEK293 cells stably expressing the wild-type (WT) isoform yielded LB-like organelles, LB-like vesicles were not readily apparent in cells expressing the ABCA3^E292V variant (Fig. 2B). By sucrose gradient fractionation and immunoblotting, a subcellular fraction band (Fig. 2C) corresponding to LB fraction but at a denser sucrose concertation level was present in ABCA3^E292V expressing cells (Fig. 2D). Together, these studies demonstrate the hypofunctional lipid transporter activity of ABCA3^E292V and extend prior findings (17, 42).

Figure 2.

In vitro-expressed ABCA3^E292V variant is functionally deficient. A: representative confocal images of A549 (top) and human alveolar type 2 (AT2) cells (bottom) 48 h following plasmid introduction of EGFP-tagged WT and the E292V mutant isoform of ABCA3 immunostained with Texas red-conjugated CD63 (top) and LAMP1 (bottom) antibodies. Bar, 5 µm. B: transmission electron microscope (TEM) images from HEK293 cell lines: parental (left) or stably expressing EGFP-tagged with either ABCA3^WT (middle) or ABCA3^E292V (right). LB-like organelles are apparent only in ABCA3^WT-expressing cells. Solid boxes are magnified underneath each image. Bar, 500 nm. C: tubes from sucrose gradient fractionation showing isolated LB band from mouse lung homogenate (2nd from left) and LB-like organelles from HEK293 cell homogenates stably expressing ABCA3^WT (3rd from left) or ABCA3^E292V (right). Strong and faint bands between 0.35 and 0.50 M sucrose fractions (arrowheads) where LBs are known to segregate are apparent in the mouse and ABCA3^WT samples, respectively, but not in parental or ABCA3^E292V samples. D: anti-EGFP immunoblot of ABCA3 primary translation (top bands) and processed (bottom bands) products from sucrose gradient fractionated samples taken from the two right tubes in C that included the upper (U), lower (L), or pellet (P) fractions showing the presence and absence of ABCA3^E292V protein band corresponding to the lower and upper fractions, respectively. ABCA3, ATP-binding cassette subfamily A, member 3; EGFP, enhanced green fluorescent protein; LBs, lamellar bodies; WT, wild type.

In Vivo Expressed ABCA3^E292V Alters Intracellular Surfactant Homeostasis

Given the limitations of in vitro models for contextual understanding of ABCA3^E292V mutations in AT2 cells in vivo, we next developed a mouse model in which the E292V missense mutation is knocked into the endogenous mouse mAbca3 locus (Fig. 3A). Litter size was normal with balanced sex distribution, and pups appeared healthy with the expected Mendelian patterns of gene transmission. Mice homozygous for ABCA3^E292V (E292V^+/+) survived into early adulthood without overt signs of respiratory distress or abnormal histological features (Supplemental Fig. S2A). Furthermore, at this early stage, neither total phospholipid (PL) content of bronchoalveolar lavage (BALF) (Fig. 3B) nor ABCA3 expression (Fig. 3C) was significantly different compared with WT control littermates. However, in contrast to WT, homozygous ABCA3 mice developed abnormally small LBs by 32 wk (Fig. 3D), which when assigned by unbiased quantitative stereological analysis revealed a reduction in LB size (Fig. 3E).

Figure 3.

Aberrant LB formation in mice carrying the E292V^+/+ variant. A: schematic illustration of the pgk-Neo⁺ (mABCA3^E292V-Neo⁺) mice showing the FRT-pgk-gb2-Neo/km-FRT cassette insert site within intron 8 in the mAbca3 locus together with the E292V point mutation in exon 9 (top) and the removal of the cassette by breeding mice with constitutively active flippase (Flp) deleter mice to produce pgk-Neo free (mABCA3^E292V) mice (bottom). B: total bronchoalveolar lavage (BAL) phospholipid (PL) content of 16-wk-old homozygous mABCA3^E292V (E292V^+/+) and WT littermates were measured; WT, n = 7; E292V^+/+, n = 6. C: mRNA levels from E292V^+/+ mice relative to WT littermates as measured by real-time quantitative PCR from at least five mice per group. D: representative electron micrographs of AT2 cells of 32-wk-old WT and E292V^+/+ mice. AT2 cell profiles in E292V^+/+ mice appear to comprise smaller lamellar bodies compared with WT controls. Boxed regions are enlarged below each image to provide enhanced resolution. E: unbiased stereological measure of LB size using volume-to-surface ratio of LB as a surrogate parameter for number-weighted mean volume of LB (V-to-S ratio) (left) and volume-weighted mean volume of LB [νV (lb) µm³] (right). *P < 0.05 vs. WT using unpaired 2-tailed t test. WT, n = 5; E292V^+/+, n = 6. F: tubes from sucrose gradient fractionation to isolate LBs from mouse lung homogenates of WT (left) and E292V^+/+ (right) mice. A band between 0.42 and 0.47 M sucrose fractions (left arrow) where LBs are known to segregate is apparent in the WT mouse sample, whereas the band from the E292V^+/+ mouse homogenate segregates below the expected level from that of the WT band (right arrow). AT2, alveolar type 2; LBs, lamellar bodies; WT, wild type.

Sucrose gradient fractionation analysis demonstrated that LB fractions from E292V^+/+ mice segregated at a denser sucrose concentration level compared with WT littermates (Fig. 3F), indicative of reduced LB lipid content that is consistent with the markedly small, abnormal LBs containing electron-dense inclusions observed by electron microscopy in Fig. 3D.

Despite this altered intracellular surfactant homeostasis, levels of the lamellar body/surfactant-associated hydrophobic proteins, surfactant protein (SP-B), and SP-C in AT2 cells (Supplemental Fig. S3A) and levels of mature SP-B and SP-C in surfactant of E292V⁺⁺ mice (Supplemental Fig. S3B) were unaffected at any of the ages examined. However, altered SP-C processing in AT2 cells was noted in 32-wk-old E292V⁺⁺ mice (Supplemental Fig. S3A).

ABCA3^E292V Mice Exhibit an Age-Dependent Macrophage-Predominant Alveolitis and Aberrant Alveolar Collagen Deposition

Total BALF cell counts were obtained at 4, 8, 16, and 32 wk postnatal age of individual mice (Fig. 4A). Although WT animals showed relatively stable BALF cell numbers, an allele-dependent alveolitis developed with progressively increased total cell counts up to 32 wk of age. Evaluation of BALF cytospin preparations indicated that this was a monocyte/macrophage-predominant population where large foamy activated macrophages appeared at 32 wk, whereas no significant phenotypic differences were noted in 8- or 16-wk-old mice (Fig. 4B). However, although we observed a wider range of variability in the BALF levels of cytokines in the E292V^+/+ mice, no significant difference was observed in E292V^+/+ mice with average ages of 16 and 32 wk compared with WT controls (Supplemental Fig. S6). These cytokines included those involved in the recruitment of monocytes and activated macrophages (MCP-1/CCL2, TARC/CCL17) and in the recruitment of neutrophils (MIP-2/CXCL2) [also can be produced by AT2 cells (44, 45)]. Hematoxylin-eosin (H&E)- and trichrome-stained sections of 32-wk-old E292V^+/+ mice also showed no gross structural abnormalities (Fig. 6B, top). Nevertheless, Picrosirius red staining revealed increased alveolar fibrillar collagen content in E292V^+/+ mice compared with WT littermate controls (Fig. 4C). The increased collagen production was further confirmed by quantitation of collagen gene biomarkers using RT-qPCR where significant increases in the messages of Col1a1 and Col3a1 that are involved in directing the production of type I and type III collagen, respectively, were noted (Fig. 6B, bottom). Moreover, quantitative stereological analysis of toluidine blue-stained sections of 32-wk-old ABCA3^E292V mice showed a wide range of variability in septal wall thickness and acinar airspace volume among E292V^+/+ mouse alveoli (Supplemental Fig. S2B).

Figure 4.

A: BALF total cell counts from 4-, 8-, 16-, and 32-wk-old E292V^−/+, E292V^+/+, and WT control littermates. *P < 0.05 vs. WT using one-way ANOVA with post hoc Tukey’s test. 4, 8, 16, 32 wk: WT, n = 10, 7, 8, 11; E292V^-/+, n = 9, 10, 9, 5; E292V^+/+; n = 10, 8, 9, 10, respectively. B: representative cytospin samples collected from 8-, 16-, and 32-wk-old E292V^+/+ mice and WT control littermates. Inset: a higher-resolution image of monocytes and macrophages. C: picrosirius red-stained lung section from 16- and 32-wk-old WT and E292V^+/+ mice (left) and dot plots of % area of alveolar collagen from transferred images to ImageJ Data software (right), each dot representing percentage of at least 10 randomly selected fields per mouse lung section (right). Bar, 50 µm. *P < 0.01 using one-way ANOVA with post hoc Tukey’s test. 16 wk, n = 5 per group; 32 wk, WT, n = 5; E292V^+/+, n =  6. BALF, bronchoalveolar lavage fluid; WT, wild type.

Age-Dependent Autophagy Disruption and Apoptotic Induction in AT2 Cells of ABCA3^E292V Mice

On the basis of altered size and number of LBs in the AT2 cells of E292V^+/+ mice (Fig. 3), coupled with prior data demonstrating the functional deficiency provoked by the ABCA3^E292V mutation (17, 42), we hypothesized that AT2 cell homeostasis is disrupted by ABCA3 via disturbance of cell quality control. Evaluation of the effect of E292V expression on 16- and/or 32-wk-old freshly isolated AT2 cells was performed by biochemical profiling for key molecular components associated with two of the major mechanisms of the cellular quality control system, unfolded protein response/ER-associated degradation/ubiquitin-proteasome system (UPR/ERAD/UPS) and macroautophagy.

We first evaluated a potential contribution of UPR/ERAD in response to the expression of E292V^+/+. Compared with WT controls, no significant difference was observed in any of the apparent ER stress signals in E292V^+/+ AT2 cells isolated at 16 and 32 wk including GRP78/Bip (Supplemental Fig. S4A) or HDJ2/HSP-40 (Supplemental Fig. S4B, middle lane), major ER chaperones central for ER protein quality control and UPR signaling pathways, or XBP1 (Supplemental Fig. S4B, top lane), a transcription factor that is upregulated during ER stress through its mRNA splicing by inositol-requiring enzyme 1 (IRE1) (46). Moreover, no significant detectable changes in total protein polyubiquitination, associated with both proteasome and autophagic degradation, were observed in these mice compared with WT controls (Supplemental Fig. S4C).

We next examined the influence of ABCA3^E292V expression on autophagy using a variety of independent steady-state measures. In contrast to the aforementioned ER stress signals, AT2 cells of E292V^+/+ mice showed significantly altered macroautophagy. LC3, a mammalian homolog of yeast Atg8, was used as a marker for developing autophagosomes. Compared with WT littermate controls, AT2 cells from 32-wk-old homozygous ABCA3^E292V mice (E292V^+/+) showed a significant increase in both nonlipidated (I) and lipidated (II) forms of LC3 (Fig. 5A, row 1, lane 4). Similarly, a marked increase in p62 (SQSTM1), an autophagic substrate that facilitates delivery of polyubiquitinated substrates to the autophagosomes through direct interaction with LC3 during autophagosome biogenesis (47–49), was observed in 32-wk-old AT2 cells of E292V^+/+ mice compared with controls (Fig. 5A, row 2, lane 4). One of the consequences of abnormal autophagy is the accumulation of organelles such as mitochondria due to impaired organelle recycling. When compared with 32-wk-old WT AT2 cell littermate controls, elevated Tom20, a constituent of the mitochondrial outer membrane, indicates an increase in overall mitochondrial biomass in AT2 cells of E292V^+/+ mice (Fig. 5A, row 4, lane 4). An increase in steady-state lysosomal mass was observed at 32 wk. Using LAMP1 as a marker (Fig. 5A, row 5, lane 4), increased expression was found consistent with disrupted autophagosome-lysosome turnover.

Figure 5.

Autophagy is disrupted in AT2 cells of E292V^+/+ mice. A: representative immunoblots from whole cell lysates of freshly isolated mouse AT2 cells showing baseline levels of autophagy-associated species (left) and band intensity of densitometry quantitation from Western blots of at least three separate experiments (right) of 16- and 32-wk-old E292V^+/+ mice and WT littermates. Each dot represents AT2 cell samples pooled from at least six mice. *P < 0.01 vs. WT using one-way ANOVA with post hoc Tukey’s test. LC3, p62, and Tom20, n = 7 per group; LAMP1, n = 6 per group. B: representative immunoblots of freshly isolated AT2 cells treated with bafilomycin A1 (Baf. A1) for 18 h showing increased levels of LC3 and p62 species at baseline and reaching optimum level at lower doses of Baf. A1 treatment in E292V^+/+ mice (bottom) compared with WT control littermates (top). C: representative immunoblots of whole cell lysates from AT2 cells treated with an autophagy inducer, Rapamycin (Rapa), alone or together with an inhibitor of autophagy, Baf. A1. D: representative immunoblots (top) and band intensity densitometer quantitation (bottom) showing levels of the apoptosis marker, activated/cleaved caspase 3, in AT2 cells of 16- and 32-wk-old E292V^+/+ mice. Each dot represents cleaved bands from AT2 cell samples pooled from at least six mice. *P < 0.05 vs. WT using unpaired 2-tailed t test. WT, n = 5; E292V^+/+, n = 6. ns, not significant. E: representative epifluorescence images of lung sections from 32-wk-old E292V^+/+ mice and WT control stained for SP-C (red) and apoptotic cells using TUNEL assay (green). Nuclei are stained with DAPI (blue). Solid boxes are magnified beneath each image to illustrate double positive (TUNEL⁺ + SP-C⁺) AT2 cells (arrows). Areas of patchy distribution of clustered TUNEL-labeled AT2 cells were noted in the lungs of 32-wk-old E292V^+/+ mice. Bar, 50 µm. Bottom: dot-plots representing % of double-positive cells per total SP-C⁺ cells per mouse from at least 10 fields (each at ×20 magnification). *P < 0.05 vs. WT using unpaired 2-tailed t test. WT, n = 5; E292V^+/+, n = 6. AT2, alveolar type 2; SP-C, surfactant protein C; TUNEL, terminal deoxynucleotidyl transferase dUTP nick end labeling.

Autophagic Flux Measurement Indicates a Late Autophagic Block in AT2 Cells of ABCA3^E292V Mice

Autophagy, downstream inhibition, and upstream induction was examined in AT2 cells of E292V^+/+ mice. AT2 cells were first treated with bafilomycin A1 (Baf. A1) to inhibit late-phase autophagy at the level of autophagosome-lysosome fusion. WT control AT2 cells showed a dose-dependent rise in levels of LC3 and p62 with Baf. A1 treatment (Fig. 5B, top). In contrast, in E292V^+/+ AT2 cells, the high level of LC3 and p62 at baseline either peaked at a lower dose (p62) or remained high and unchanged with increased doses of Baf. A1 (LC3) (Fig. 5B, bottom). Rapamycin, an inhibitor of the mammalian target of rapamycin (mTOR), shown to activate autophagy and alter LC3 turnover (50) followed by addition of Baf. A1 was next used to treat AT2 cells. As demonstrated in Fig. 5C, rapamycin treatment resulted in a decrease of LC3 isoforms and p62 in 32-wk-old WT AT2 cells, consistent with enhanced utilization and degradation of a rate-limiting cellular pool of LC3 and p62 proteins. In contrast, and consistent with the result shown in Fig. 5B, the high basal level of LC3 at 32 wk in E292V^+/+ AT2 cells could neither be reduced by rapamycin nor could it be further enhanced by Baf. A1 (Fig. 5C, top two lanes), whereas addition of rapamycin had no effect in reducing p62 expression (Fig. 5C, third lane from top), suggesting a late block in autophagy. Despite the absence of detectable ER stress involvement (Supplemental Fig. S4), increased cell death signals were observed in 32-wk-old E292V^+/+ AT2 cells. Compared with WT littermate controls, there was significantly augmented cell death in the 32-wk-old E292V⁺⁺ AT2 cells as demonstrated by the expression of cleaved caspase 3 (Fig. 5D) and by fragmented DNA detection via TUNEL assay (Fig. 5E), suggesting ER stress-independent and impaired autophagy-dependent induction of apoptosis as reported previously (51, 52).

Injury/Repair Dysregulation in the ABCA3^E292V Lung

Although impairment of AT2 cell autophagy and alveolar inflammation were accompanied by some features of fibrotic remodeling, we hypothesized that the observed cellular phenotype would enhance the propensity of these mice to develop aberrant lung injury/repair responses to exogenous “second hits.” To assess this, intratracheal (IT) instillation of bleomycin, a commonly used model of subacute experimental lung injury and fibrosis (53–55), was performed in E292V^+/+ mice. A significant age-dependent morbidity and mortality was observed in older IT bleomycin-treated E292V^+/+ mice. Compared with WT littermate controls, bleomycin instillation did not show any apparent injury/repair differences in 16-wk-old E292V^+/+ mice (data not shown). However, although no difference was observed in BALF cell counts between E292V^+/+ and WT littermate controls (Fig. 6E), relatively rapid and significant weight loss differences were noted in 32-wk-old E292V^+/+ mice (Fig. 6A). Moreover, results from H&E- and trichrome-stained lung sections of 32-wk-old E292V^+/+ mice 21 days following bleomycin appeared to suggest augmented fibrosis throughout the E292V^+/+ mouse lung as compared with WT, where areas of patchy fibrotic lesion were observed in predominantly normal looking lung (Fig. 6C). Quantitative determination by Picrosirius red staining substantiated the relative increase in alveolar fibrillar collagen content in E292V^+/+ mice compared with WT controls (Fig. 6D). The aberrant lung remodeling was accompanied by hyperplastic AT2 cells (Fig. 6E and Supplemental Fig. S5) to suggest aberrant lung injury/repair processes. Consistent with these findings, pulmonary function tests revealed restrictive lung physiology in the bleomycin-treated E292V^+/+ mice (Fig. 6F). Finally, the vulnerability to lung injury caused by this mutation was further manifested at a higher dose of bleomycin treatment where lethality was significantly increased in E292V^+/+ mice compared with WT controls (Fig. 6G).

DISCUSSION

Although the prevalence of the ABCA3^E292V mutation is high in the general population and further enriched in certain ethnic groups, its precise role in the pathogenesis of lung disease is incompletely defined. We describe a clinical case of a homozygous E292V patient with progressive pulmonary dysfunction who underwent lung transplantation at age 47 yr due to respiratory failure, 22 years after becoming symptomatic. Our in vitro data further defined a functional deficiency in the E292V variant extending previous reports (41, 42, 56). In vivo, the initial examination of mice constitutively expressing the mutant transporter showed no apparent baseline alteration in perinatal survival, development, bronchoalveolar lavage (BAL) phospholipid level, Abca3 RNA expression, or alveolar morphology. However, despite maintaining extracellular surfactant homeostasis, data in the present study suggest that E292V alters/disrupts AT2 cell homeostasis, as subsequent biochemical, immunological, histological, and injury/remodeling data from these mice revealed notable changes in key AT2 cell quality control pathways, increased lung inflammation, enhanced alveolar fibrillary collagen deposition, and an increased vulnerability to exogenous second hits.

Consistent with previous reports, our in vitro models showed that although the ABCA3^E292V variant is trafficked normally, localizing in lysosomal-like organelles in A549 and HEK cells and in LBs of AT2 cells, it is hypofunctional as a lipid transporter (Fig. 2). The use of cell lines demonstrated the hypofunctional activity of ABCA3^E292V in a normally trafficked transporter, and the phenotypical alteration by ABCA3^E292V expression in these cell lines was not expected since, unlike AT2 cells, they do not require either lamellar bodies or surfactant production to maintain cellular homeostasis. Characteristics of normal protein trafficking but reduced ATPase activity have been reported for several other ABCA3 mutations (41, 42, 56). Moreover, a study conducted using ABCA3 mutant proteins associated with pediatric interstitial lung disease (pILD) including E292V, E690K, and T1114M has shown that although all three transporters are normally trafficked, the lipid transport function of E690K and T114M is markedly impaired, whereas E292V function is moderately preserved (17).

Despite this, the ABCA3^E292V mouse model demonstrates ultrastructural changes manifested by small LBs containing electron-dense inclusions (Fig. 3). This characteristic morphology is also observed in LBs of infants and children carrying many of the ABCA3 mutations (7, 14, 57–60), highlighting one of the downstream consequences of ABCA3 mutations that produce dysfunctional or hypomorphic transporters. We have also previously reported a similar outcome of small, electron-dense LBs in AT2 cells of a mouse model of surfactant deficiency caused by dysregulation of the Abca3 gene (20). Together, these findings further substantiate the essential role of ABCA3 in the selective and measured transport of specific phospholipid species into LBs that is necessary for the appropriate composition and function of alveolar surfactant (61).

Although the E292V mutant protein is trafficked normally and thus does not induce ER stress (Supplemental Fig. S3), as has been reported for other ABCA3 mutations (62), the present study demonstrates that E292V expression does produce a distinct ABCA3 mutation-related cellular phenotype primarily marked by alteration of a key cell quality control process, autophagy. Perturbations in autophagy can be assessed by the steady-state number of autophagosomes and/or alterations in the autophagosome-associated form of vesicle-associated proteins LC3 and p62, involved in the autophagosomal as well as proteasomal degradation of ubiquitinated proteins (50, 63). We found age-dependent accumulation of autophagosomes in E292V^+/+ AT2 cells, as assessed by elevated LC3 and p62 levels (Fig. 5A). Because both synthesis and degradation affect steady-state LC3 and p62 levels, we determined autophagic flux using Baf. A1, which prevents lysosomal acidification and promotes accumulation of LC3 and p62 by inhibiting their degradation. E292V^+/+ cells treated with Baf. A1 exhibited enhanced sensitivity to the reagent in E292V^+/+ cells compared with the WT control (Fig. 5B). Moreover, treatment with rapamycin, an autophagy inducer through inhibition of mTOR, resulted in a normal induction of LC3 and p62 turnover in WT AT2 cells, whereas no changes in LC3 isoforms were observed in E292V^+/+ AT2 cells (Fig. 5C). In total, these data are consistent with decreased autophagosome maturation/degradation (a late block in autophagy).

The age-dependent aberrant autophagic function appears to correlate with the slow progression of phenotypic abnormalities observed in the aging E292V^+/+ mice together with the moderately preserved function of the E292V transporter. Baseline biochemical and histological evaluation of AT2 cells or lung showed no detectable abnormalities at 16 wk (Fig. 5A and Supplemental Fig. S2A). Even after the mice were challenged via lung instillation of bleomycin, no apparent injury/repair differences were observed between the lungs of E292V^+/+ and WT littermates at this age (unpublished observation). Likewise, there were no significant differences observed in BALF cytokine levels in 16- and 32-wk-old mice (Supplemental Fig. S6). This is in contrast to what we have previously reported in Sftpc mouse models, where conditional expression was used to produce an extreme phenotype constituting an acute inflammatory phase that included a robust cytokine activity within the first few days of induction (23, 25, 64). One difference between these two mouse models is that although pathology is readily recognized in the conditionally expressed SP-C studies and, therefore, mostly conducted in a 3–28-day experimental window, the E292V^+/+ mouse model takes 5–10 times longer to develop pathology. Thus, the E292V^+/+ would potentially have at least a 16-wk experimental window to examine changes in cytokine levels. Although it is beyond the scope of the present study, a more detailed ontogeny of expression of potential candidates by super multiplex or an unbiased proteomics approach may decipher the cytokine profile of this mouse model. However, the E292V^+/+ phenotype appears to be more similar to the published telomerase mouse models that clearly showed that AT2 telomere disruption produces fibrosis, senescence, inflammatory cells, and bleomycin sensitivity but did not establish a clear link with specific cytokines (44, 65).

At 32 wk, mild but significant baseline changes in the lung appeared, including alveolitis (Fig. 4B) and fibrotic remodeling (Fig. 4C and Fig. 6B, bottom), and the mice developed increased susceptibility to lung injury (Fig. 6). This relatively late development of lung abnormities is likely the by-product of a slow deterioration of AT2 cell function possibly attributable to both mild surfactant insufficiency and altered autophagy. As autophagy is considered one of the most critical cellular homeostatic processes (66), an alteration of this pathway, however gradual, could modify normal cellular function, including contributing to the lung disease phenotype presented in the current study.

Thus, the age-dependent defect in the autophagy arm of the AT2 cell quality control system may play a critical role in the development and progression of lung pathogenesis. Studies have revealed characteristics of defective autophagy in AT2 cells of human patients with Hermansky–Pudlak syndrome (HPS) as well as in mouse models of HPS (67). Moreover, we have previously reported the deleterious effect in cell lines of an SFTPC mutation (SP-C^I73T) on cellular autophagy that induces a block in autophagy-dependent proteostasis and phenocopies the ultrastructural changes observed in AT2 cells from a lung biopsy of a patient suffering from DPLD (68). Our subsequent proof-of-concept study using conditional expression in a mouse model revealed the capability of SP-C^I73T to elicit spontaneous lung fibrosis in vivo (23). These studies together with our present findings suggest that disruption of AT2 cell quality control mechanisms, including autophagy, represents a critical contributor to the pathogenesis of DPLD either as a primary driver or as a modifier gene.

Moreover, the more moderate lung pathology elicited by E292V^+/+ expression in the present study is consistent with a previous report that identified heterozygous ABCA3 mutations in severely affected infants with the SP-C^I73T mutation (where two of the three infants carry the E292V mutation) (8). The study indicated that the inheritance of these mutations from disease-free parents supports the hypothesis that ABCA3 acts as a modifier gene for the phenotypic features related to an SFTPC mutation.

Interestingly, although LC3 and p62 steady-state levels were increased, a corresponding increase in baseline total polyubiquitination was not observed in the E292V^+/+ AT2 cells (Supplemental Fig. S3C). p62 serves as a linker between polyubiquitinated proteins and autophagy-mediated degradation, as it facilitates delivery of polyubiquitinated substrates to the autophagosome through direct interaction with LC3 during autophagosome biogenesis (47–49). p62 is also involved in the proteasomal degradation of ubiquitinated proteins (63). Therefore, the absence of a noticeable increase in total polyubiquitination in E292V^+/+ AT2 cells is likely due to the inherent redundancy in the cellular proteostasis system (65), wherein E292V expression-dependent polyubiquitination could be masked by robust primary or compensatory proteasome degradation. Recent mouse modeling of AT2 cell quality control by knockdown of either proteasome or macroautophagy suggests a primary role of the 26S proteasome in proteostasis, whereas autophagy appears to be critical for organellar homeostasis (69, 70).

Our demonstration of elevated levels of Tom20, an outer mitochondrial receptor complex responsible for the recognition and translocation of cytosolic mitochondrial preproteins, in E292V^+/+ AT2 cells implies an increase in mitochondrial biomass (Fig. 5A, row 4) and suggests an accumulation of dysfunctional mitochondria. One of the consequences of a block in autophagy is the disruption of mitochondrial turnover by selective degradation of dysfunctional mitochondria by mitophagy, a key quality control mechanism for bulk degradation of mitochondria. Indeed, mitophagy contributes to mitochondrial quality control not only by removing damaged mitochondria but also by promoting biosynthesis of new mitochondria (71). It has been shown that there is cross talk between the mitophagy pathway and mitochondrial biosynthetic pathway (72). Moreover, as demonstrated in Fig. 5D, an age-dependent increase in cleaved caspase 3 levels suggests a possible link between dysfunctional mitochondria and apoptosis. That is, in the absence of efficient mitochondrial turnover, the resultant accumulation of damaged mitochondria could lead to increased membrane permeability causing activation of the cell death pathway through releases of proapoptotic molecules such as cytochrome c and apoptosis-inducing factor (AIF) (73). Further studies are required to determine whether any or all of the factors including cytochrome c, AIF, and cellular stress due to surfactant insufficiency contribute to the activation of caspase 3 in AT2 cells of E292V^+/+ mice.

Although at baseline, we observed no apparent histological disruption of the alveolar structure at any timepoints examined (Supplemental Fig. S2A and Fig. 6B), older E292V mice exhibited increases in inflammatory cells (Fig. 4A) composed of abnormally large macrophages (Fig. 4B), accompanied by augmented alveolar collagen (Fig. 4C and Fig. 6B, bottom), which suggested an abnormality in lung injury/repair. Such defects were further demonstrated in Fig. 6 by age-dependent enhanced morbidity in bleomycin-treated E292V^+/+ mice, indicating a scope of severity extending to a total failure of lung function (Fig. 6A and G). E292V^+/+ mice had increased collagen deposition, exaggerated lung fibrosis (Fig. 6C and D) accompanied by hyperplastic AT2 cells (Fig. 6E and Supplemental Fig. S5), and diminished pulmonary function (Fig. 6F) compared with wild-type controls. These results suggest that E292V^+/+ expression-related surfactant insufficiency coupled with altered AT2 cell autophagy leaves the lung vulnerable to external injury and hampers its ability to adequately carry out reparative functions. The efficiency of normal injury/repair function in AT2 cells has been clearly demonstrated in a recent study of a mouse model of AT2 cell injury caused by conditional deletion of the Abca3 gene (74). Previous ABCA3 mouse models with either altered function (20) or haploinsufficiency (mAbca3^+/−) (1, 3, 4), both characterized by decreased lung phospholipid levels, have also shown susceptibilities to either bleomycin- (20) or hyperoxia-induced and mechanical ventilation-induced (75) lung injury, respectively. Therefore, although the level of the effect from either altered autophagy or surfactant insufficiency in the E292V^+/+ mice is yet to be determined, for gene mutations associated with variable disease presentations, including those observed in patients with SFTPC, SFTPA, MUC5B, or ABCA3, multiple hits in the form of coexpressed genetic modifiers and/or environmental factors are likely necessary for augmenting disease severity or inducing fibrotic lung pathology (12, 76, 77).

Although considerable progress has been made in elucidating the critical role of ABCA3 in pulmonary phospholipid homeostasis, surfactant function, and respiratory distress syndromes, understanding its role in the pathogenesis of chILD and DPLD/IPF remains challenging. Genetic and environmental factors have been implicated as contributors to phenotypic variation in ABCA3-related disorders and studies on infants carrying both ABCA3 and SFTPC mutations, implying that ABCA3 may play a role as a modifier gene (described above). Dissection of the molecular signatures of mutations such as E292V that cause cellular/lung alteration, including the quality control systems recruited to sustain cellular homeostasis, enhances our understanding of disease etiology and promises to contribute to the development of new treatment strategies. Animal models constitutively expressing mutant proteins (as in the present study) provide an invaluable tool in this endeavor.

SUPPLEMENTAL DATA

Supplemental Figs. S1–S6: https://doi.org/10.6084/m9.figshare.14527098.

GRANTS

This work is supported by the National Institutes of Health (HL129150 to S. Mulugeta, HL145408 to M.F. Beers, NIH K08 and HL 150226 to J. Katzen, and NIH 2 T32 HL007586-36 to L. Rodriguez) and the Department of Veterans Affairs (VA Merit Award BX001176 to M.F. Beers).

DISCLOSURES

M.F. Beers is an established investigator of the Pulmonary Fibrosis Foundation. None of the other authors has any conflicts of interest, financial or otherwise, to disclose.

AUTHOR CONTRIBUTIONS

J.K., M.F.B., and S.M. conceived and designed research; Y.T., L.K., M.Z., L.R.R., A.M., A.V., and , S.M. performed experiments; Y.T., J.W., L.K., M.Z., L.R.R., F.V.W., J.K., M.O., A.H., M.F.B., and S.M. analyzed data; Y.T., J.W., L.K., M.Z., F.V.W., J.K., M.O., A.H., M.F.B., and S.M. interpreted results of experiments; Y.T., J.W., L.K., M.Z., and S.M. prepared figures; J.W. and S.M. drafted manuscript; Y.T., J.W., L.K., J.K., M.O., M.F.B., and S.M. edited and revised manuscript; Y.T., J.W., L.K., M.Z., L.R.R., A.M., F.V.W., A.V., J.K., M.O., A.H., M.F.B., and S.M. approved final version of manuscript.