Proteasome dysfunction in alveolar type 2 epithelial cells is associated with acute respiratory distress syndrome

Proteasomes are a critical component of quality control that regulate turnover of short-lived, unfolded, and misfolded proteins. Proteasome activity has been therapeutically targeted and considered as a treatment option for several chronic lung disorders including pulmonary fibrosis. Although pharmacologic inhibition of proteasome activity effectively prevents the transformation of fibroblasts to myofibroblasts, the effect on alveolar type 2 (AT2) epithelial cells is not clear. To address this knowledge gap, we generated a genetic model in which a proteasome subunit, RPT3, which promotes assembly of active 26S proteasome, was conditionally deleted in AT2 cells of mice. Partial deletion of RPT3 resulted in 26S proteasome dysfunction, leading to augmented cell stress and cell death. Acute loss of AT2 cells resulted in depletion of alveolar surfactant, disruption of the alveolar epithelial barrier and, ultimately, lethal acute respiratory distress syndrome (ARDS). This study underscores importance of proteasome function in maintenance of AT2 cell homeostasis and supports the need to further investigate the role of proteasome dysfunction in ARDS pathogenesis.

weight after 4 days of treatment but began to gain weight after day 5 ( Supplementary Fig. S1a). Daily assessment of health indicated that RPT3 AT2Δ/Δ and Cre mice gained weight after transition to regular chow and were active. However, on day 10, RPT3 AT2Δ/Δ mice began to lose weight ( Supplementary Fig. S1a). On day 11, all RPT3 AT2Δ/Δ mice experienced a precipitous decline in health and activity: 53% of RPT3 AT2Δ/Δ mice (n = 17/32) lost 7.5% of body weight from day 10 (12% loss from day 0) leading to morbidity and death within 3-4 hours, and 47% of RPT3 AT2Δ/Δ mice (n = 15/32) lost greater than 20% body weight and were immediately euthanized. In contrast to RPT3 AT2Δ/Δ mice, Cre mice maintained on tamoxifen chow for 7 days were active and healthy on day 11 and continued to gain weight ( Supplementary Fig. S1a).
To identify a deletion strategy that did not result in acute morbidity, mice were treated with tamoxifen for shorter periods of time. Mice were fed tamoxifen chow for 1, 3, 4 or 5 days, transitioned to regular diet, monitored daily, and surviving mice euthanized 3.5 weeks after removal of tamoxifen chow ( Supplementary Fig. S1c). Tamoxifen treatment for 5 days resulted in lethality on day 11 of the study (n = 2/3), similar to the 7-day tamoxifen treatment regimen; therefore, recombination efficiency at the Psmc4 locus was assessed following a 4-day treatment regimen. Quantitative PCR and Western blot analyses of isolated AT2 cells detected no significant changes in Psmc4 mRNA or RPT3 protein and all mice survived to day 35 when the experiment was terminated ( Supplementary Fig. S1e,f). Since recombination was minimal after 4 days of tamoxifen treatment, all further studies were conducted using the 7-day tamoxifen treatment regimen.
AT2 cell-specific deletion of RPT3 leads to disruption of the alveolar epithelial barrier. RPT3 deficiency resulted in an acute and rapid loss of AT2 cells, followed by sudden onset of respiratory distress and death on day 11. Paradoxically, analysis of H&E stained lung sections of RPT3 AT2Δ/Δ mice prior to presentation of symptoms on day 11 revealed normal lung structure with only very mild inflammation (Supplementary Fig. S2b-d). Consistent with histological analyses, alveolar structure appeared relatively unaffected when examined by scanning electron microscopy (SEM) (Fig. 3a-d). However, transmission electron microscopy (TEM) identified cell debris (Fig. 3g,h), damaged alveolar type 1 (AT1) epithelial cells with pronounced membrane blebbing ( Fig. 3k-l) and alveoli with vacuolated AT2 cells (Fig. 3o). Additionally, AT2 cells from RPT3 AT2Δ/Δ mice contained aggregates (Fig. 3o) and electron dense lamellar bodies (Fig. 3p) compared to control AT2 cells (Sftpc WT/CreER :RPT3 F/F mice without tamoxifen and Sftpc WT/CreER mice with tamoxifen treatment). Consistent with TEM analyses, lung sections stained with T1α (Podoplanin) revealed damaged AT1 cells on day 11, including cells that appeared distended, with T1α + debris in the airspaces ( Supplementary Fig. S3a). Importantly, ultrastructural changes in AT1 cells were detected as early as day 9 (Fig. 3k) coincident with a 60% decrease in the frequency of AT2 cells (Fig. 2c). Structural changes were more pronounced in the alveolar epithelium compared to the endothelium; significant qualitative changes were not observed in endothelial cells by TEM analyses (Fig. 3l) or by confocal imaging of lung sections stained with EMCN (Endomucin) (Supplementary Fig. S3b). Collectively, these data indicate that RPT3 deficiency leads to AT2 cell injury and disruption of the alveolar epithelial barrier.
To assess the impact of epithelial damage on barrier function, epithelial barrier permeability was measured on day 11, prior to presentation of any symptoms. Mice were intravenously injected with albumin conjugated to FITC, followed by measurement of FITC fluorescence in BALF. RPT3 AT2Δ/Δ mice demonstrated a 3.5-fold increase in FITC fluorescence compared to Sftpc WT/CreER :RPT3 F/F mice (Sftpc WT/CreER :RPT3 F/F = 0.23 ± 0.1, Sftpc WT/CreER = 0.23 ± 0.1, RPT3 AT2Δ/Δ = 0.8 ± 0.4), indicating an increase in barrier permeability in response to RPT3 deficiency ( Supplementary Fig. S4a). Consistent with increased barrier permeability, a 4.4-fold increase in protein concentration was observed in BALF recovered from an independent group of RPT3 AT2Δ/Δ mice compared to Sftpc WT/CreER :RPT3 F/F mice (Sftpc WT/CreER :RPT3 F/F = 0.5 ± 0.1, Sftpc WT/CreER = 0.74 ± 0.4, RPT3 AT2Δ/Δ = 2.2 ± 0.96) ( Supplementary Fig. S4b). Although the latter group of RPT3 AT2Δ/Δ mice demonstrated a significant increase in neutrophil frequency in BALF ( Supplementary Fig. S4c), the total number of immune cells remained unchanged ( Supplementary Fig. S4d), consistent with the mild inflammation observed in histological analyses Adaptive changes in response to RPT3 deficiency. To identify transcriptomic changes in response to partial deletion of RPT3, RNA sequencing was performed on AT2 cells isolated from RPT3 AT2Δ/Δ and Cre mice on day 9 after tamoxifen treatment (days 1-7); day 9 was chosen to identify gene expression changes associated with a 60% decrease in AT2 cell frequency (Fig. 2c). In response to RPT3 deficiency, 6569 genes were differentially expressed, with 53.4% genes (n = 3509/6569) upregulated and 46.6% genes (3060/6569) downregulated ( Supplementary Fig. S5a). Gene ontology (GO) analysis revealed significant enrichment of genes associated with catabolic and metabolic processes, ubiquitin-dependent proteolysis, cell stress, cell death and apoptosis (Supplementary Table S1). Differential expression of 175 proteasome-associated genes was observed in response to RPT3 deficiency ( Supplementary Fig. S5b), with 83.1% genes (n = 148/178) upregulated and 16.9% genes (n = 30/178) downregulated. Upregulated genes included majority of the proteasome subunits and regulators including Psmc4 (RPT3) (Supplementary Table S2 and Fig. S5c,d). Analysis of RNA sequencing track data and qPCR on isolated AT2 cells using primer-probe sets directed to exons 3-4 and 9-11 revealed that increase in Psmc4 expression was due to increased formation of a stable, truncated mRNA (Supplementary Figs. S1b and S5e). RPT3 deletion leads to accumulation of proteasome substrates in AT2 cells. The 26S proteasome comprises a 20S catalytic core and a 19S regulatory particle which caps one or both ends of the 20S complex. The 19S complex recognizes ubiquitinated substrates targeted for degradation by the 20S core. RPT3 is an AAA + -ATPase subunit of the 19S complex and is required for assembly of the 26S proteasome. Native gel electrophoresis of isolated AT2 cells was performed to determine the effect of increased expression of genes encoding both 19S and 20S proteasome subunits (Supplementary Table S2). Immunoblotting of native gels demonstrated an increase in 20S α1-7 subunits and the 19S subunit RPT5 (Fig. 4a), consistent with differential gene expression analyses ( Supplementary Fig. S5c). Accumulation of 20S proteasome content was accompanied by an increase in 20S chymotrypsin-like activity in AT2 cells isolated from RPT3 AT2Δ/Δ mice on day 9 compared to Sftpc WT/CreER :RPT3 F/F mice (Fig. 4a). However, SDS-PAGE of AT2 cells isolated from RPT3 AT2Δ/Δ mice revealed a dramatic accumulation of polyubiquitinated and K48-linked polyubiquitinated substrates (Fig. 4b). Collectively, these data indicate that although RPT3 deficiency is associated with an increase in 19S and 20S proteasome subunits, 26S proteasome proteolytic activity is impaired leading to reduced protein turnover and increased cellular aggregate load. www.nature.com/scientificreports www.nature.com/scientificreports/ Proteasome dysfunction results in increased cell stress and death. Proteasome dysfunction triggers stress responses that allow cells to adapt to aggregate load or cause apoptosis 24 . AT2 cell-specific deletion of RPT3 resulted in global cell-stress response and was associated with upregulation of genes that encode heat shock proteins, ER chaperones and co-chaperones, and integrated stress response proteins (Fig. 5a); a comprehensive list of genes is provided in supplementary file 2. Western blot analyses of AT2 cells isolated on day 9 from RPT3 AT2Δ/Δ mice revealed an increase in cytosolic cell stress chaperone HSP70, ER stress chaperone BiP and downstream effectors of the integrated stress response, eIF2α, ATF4 and GADD34, compared to Sftpc WT/CreER : RPT3 F/F mice (Fig. 5b).
Autophagic activity is saturated in response to RPT3 deletion. Increasing evidence suggests that autophagy can function as a backup system to help relieve proteasome overload [25][26][27] . To determine if basal autophagy was altered in response to proteasome dysfunction, LC3B levels were assessed by Western blotting. Despite upregulation of genes associated with the autophagy-lysosome pathway, (n = 113/176) ( Supplementary  Fig. S7a), LC3BI and lipidated LC3BII levels were not significantly altered in AT2 cells isolated from RPT3 AT2Δ/Δ mice compared to Sftpc WT/CreER :RPT3 F/F mice ( Supplementary Fig. S7b-d). LC3B was primarily detected as the lipidated LC3BII form, consistent with previous data from isolated AT2 cells 28 . Further, expression of Becn1, an upstream regulator of autophagy 29 , was unaltered ( Supplementary Fig. S7e), suggesting that deletion of RPT3 did not promote augmented autophagosome biogenesis at steady state. The latter hypothesis was tested by measuring autophagic flux in vitro using AT2 cells isolated from control mice (Sftpc WT/CreER :RPT3 F/F or tamoxifen treated Sftpc WT/CreER mice) and RPT3 AT2Δ/Δ mice. In response to either Bafilomycin A1 or chloroquine, a significant increase was observed in LC3BII levels in control AT2 cells (DMSO: 0.524 ± 0.05, 15 nM Bafilomycin A1: 1.2 ± 0.11, 40 μM chloroquine: 1.5 ± 0.3) consistent with active autophagic flux (Fig. 6a,b). In contrast, LC3BII levels were already elevated in DMSO-treated AT2 cells isolated from RPT3 AT2Δ/Δ mice (1.2 ± 0.32), with no further increase in response to Bafilomycin A1 or chloroquine (15 nM Bafilomycin A1: 1.9 ± 0.73, 40 μM chloroquine: 1.6 ± 0.4) (Fig. 6a,b). Collectively, these data suggest that in response to deletion of RPT3 flux through the autophagy-lysosome system is near maximum, with little to no change in clearance of polyubiquitinated substrates.

Discussion
RPT3 promotes assembly of the 19S particle with the 20S catalytic core to form active 26S proteasome. Partial deletion of RPT3 in AT2 cells resulted in dramatic accumulation of polyubiquitinated substrates that overwhelmed cellular disposal pathways leading to elevated cell stress and, ultimately, AT2 cell death. Rapid and extensive loss of AT2 cells was accompanied by depletion of alveolar surfactant and disruption of the alveolar epithelial barrier; the inability to rapidly re-epithelialize damaged alveoli resulted in lethal ARDS.
Formation of the 26S proteasome is dependent upon interactions between the C-terminal domain of 19S RPT subunits with 20S α subunits 32 . RPT subunits bind and hydrolyze ATP, which is required for substrate unfolding and 26S proteasome assembly and activation 13 ; deletion of exons 7-11 of the Psmc4 gene (RPT3) disrupts the AAA + -ATPase domain 20 . 26S proteasomes harboring ATP binding mutations in single RPT subunits have little effect on ATP hydrolyzing ability of the other five normal RPT subunits and can degrade small peptides, such as Suc-LLVY-AMC; however, these mutant proteasomes are much less efficient in degrading polyubiquitinated proteins 33 . RPT3 deletion resulted in accumulation of polyubiquitinated substrates, suggesting an inability to stimulate proteolytic activity of the 26S proteasome. Despite an adaptive increase in free 20S and 19S complexes, no apparent change was observed in 26S content or activity suggesting that loss of RPT3 resulted in de novo formation of few or unstable 26S proteasomes. Additionally, it is possible that aggregated proteins inhibited extant 26S proteasome function leading to further accumulation of potentially cytotoxic substrates.
p62 sequesters polyubiquitinated substrates into inclusions to protect the cell from cytotoxic aggregates; furthermore, p62 can deliver polyubiquitinated proteins to the autophagic machinery for clearance via the lysosome 34,35 . RPT3 deletion resulted in significant upregulation of p62 protein and mRNA, prior to accumulation of polyubiquitinated substrates. However, defective clearance of polyubiquitinated proteins despite increased lysosomal degradation suggests that sequestration of p62 within aggregates may impair substrate delivery to the lysosome resulting in augmented accumulation. www.nature.com/scientificreports www.nature.com/scientificreports/ and p62 (c) levels from three independent flux experiments. Samples and immunoblots were processed in parallel. Control included Sftpc WT/CreER :RPT3 F/F (RPT3 F/F ) mice without tamoxifen treatment or Sftpc WT/CreER mice that were fed tamoxifen chow for 7 days. *p < 0.05, **p < 0.01, ***p < 0.001 by two-way ANOVA with Sidak's multiple comparison test. (d) Western blot analysis and densitometry (e) of p62 in 30 μg of AT2 cell lysates separated by SDS-PAGE. (f) Quantitative PCR for Sqstm1 (p62) in isolated AT2 cells. RQ: relative quantitation. *p < 0.05, **p < 0.01, ***p < 0.001 by one-way ANOVA with Tukey's multiple comparison test. (g) Representative single focal plane image of lung sections stained for AT2 cells with proSP-C and ABCA3, and p62. Insets show individual channels for proSP-C, ABCA3, and p62. Scale bars = 10 μm. Confocal z-stacks were deconvoluted to obtain final images (details in methods). Controls included both RPT3 F/F mice (Sftpc WT/CreER :RPT3 F/F mice without tamoxifen treatment) and Cre mice that were fed tamoxifen chow for 7 days. Full-length blots are presented in Supplementary Fig. 9 www.nature.com/scientificreports www.nature.com/scientificreports/ AT2 cell-specific deletion of RPT3 resulted in coincident increases in ubiquitinated substrate load, cell stress and apoptosis, consistent with a recent study published by Kitajima et al. 36 . These results are also consistent with the finding that treatment of cultured mouse AT2 cells with proteasome inhibitors resulted in dose-dependent cytotoxicity 7 . It is not clear if accumulation of ubiquitinated proteins is directly cytotoxic and/or if sequestration of regulatory proteins into aggregates leads to collapse of key cellular processes 37 . It is also possible that failure to degrade ubiquitinated substrates leads to depletion of essential amino acids resulting in induction of the integrated stress response 38 . Overall, the precise sequence of events between substrate accumulation and AT2 cell death remain unclear.
AT2 cells are facultative progenitors in the adult lung and respond to injury or loss by self-renewal and differentiation to AT1 cells 39 . RPT3 AT2Δ/Δ mice demonstrated respiratory failure following rapid loss of 90% AT2 cells over a 4-day period, with no evident compensatory proliferation. Deletion of RPT3 from muscle satellite cells inhibited proliferation in a p53-dependent manner and increased apoptosis 36 . Consistent with this report, a significant increase in Trp53 and its associated pathways was observed in response to deletion of RPT3 from AT2 cells (supplementary file 2). Notably during the cell cycle, 26S proteasome is post-translationally modified, predominantly by phosphorylation of 19S and 20S subunits; inhibition of RPT3 phosphorylation resulted in 26S proteasome dysfunction, dysregulated degradation of cell cycle regulators and delayed cell cycle progression 40 . Collectively, these results support a model in which RPT3 deficiency inhibits proliferation of all or a subset of AT2 progenitors 41 , resulting in continual cell loss. The lethal ARDS-like phenotype observed in response to proteasome dysfunction in RPT3 AT2Δ/Δ mice is consistent with previous reports from AT2 cell-ablation models. Depletion of 80% AT2 cells over a 10-day period resulted in respiratory failure in SPC-TK mice 42 . Similarly, administration of diphtheria toxin to LysM-DTR mice resulted in lethality associated with depletion of AT2 cells 43 .
AT2-cell specific deletion of RPT3 resulted in acute morbidity and mortality associated with a 50% decrease in Psmc4 (RPT3) expression. The lethal ARDS-like phenotype was associated with injury of both AT2 and AT1 cells, the latter likely secondary to loss of junctional complexes that anchor AT1 cells to AT2 cells; indeed, differential gene expression and GO analyses revealed downregulation of genes encoding tight junction proteins (supplementary file 2). Injury to the alveolar epithelial barrier was extensive compared to the underlying capillary endothelium. Alveolar surfactant pool size in RPT3 AT2Δ/Δ mice was decreased by 90% immediately prior to presentation of respiratory symptoms. A similar decrease in surfactant pool size was reported in response to deletion of ABCA3 in AT2 cells 44 ; however, respiratory failure was delayed by more than a week in ABCA3 knockout mice. These findings provide further support for the hypothesis that extensive AT2 cell death and loss of barrier integrity are the primary drivers of acute respiratory failure in RPT3 AT2Δ/Δ mice; it is also likely that elevated surface tension arising from surfactant deficiency exacerbated AT2 cell injury and subsequent barrier dysfunction.
To date few studies have reported an association between proteasome dysfunction and ARDS: ubiquitin positive inclusion bodies were detected in alveolar epithelial cells of patients with diffuse alveolar damage 45 , and proteolytically active 20S proteasomes were released from damaged AT2 cells and recovered from BALF of patients with severe ARDS 46,47 . A recent case report identified complete alveolar denudation in two patients with diffuse alveolar damage and severe ARDS consistent with failure of AT2 cell-mediated lung repair 48 . In the current study, AT2 cell-specific deletion of Psmc4 (RPT3) by 50% was sufficient to impair 26S proteasome function resulting in rapid AT2 cell apoptosis and lethal ARDS associated with an absence of AT2 cell proliferation and lung repair. Although the severe phenotype precludes the use of RPT3 AT2Δ/Δ mice as a clinical model of ARDS, this study underscores the importance of proteasome function in maintenance of AT2 cell homeostasis and supports the need to further investigate the role of proteasome dysfunction in ARDS pathogenesis.

Methods
Mice. RPT3 F/F mice on C57/Bl6 background were obtained from The Center for Animal Resources and Development (CARD), Kumamoto University, Japan 20 and crossed with the Sftpc CreERT2 mice [on a mixed C57:129 genetic background 49 ]. Genotyping was performed from tail biopsies; RPT3 F/F mice were genotyped with allele specific primers as described previously by Tashiro et al. 20 and primers recommended by the Jackson Laboratory were used to distinguish between heterozygosity and homozygosity for the Cre allele. 8-12-week old female and male Sftpc WT/CreER :RPT3 F/F and Sftpc WT/CreER control mice were provided ad libitum access to tamoxifen chow (400 mg tamoxifen citrate, Envigo). Mice were housed in a pathogen-free barrier facility. All procedures were performed under protocols (IACUC2015-0073: TEW) approved by the Institutional Animal Care and Use Committee of Cincinnati Children's Hospital Medical Center in accordance with National Institutes of Health guidelines. Mice were euthanized when they demonstrated any of the following symptoms: kyphosis, severe lethargy and inactivity as observed by a lack of response when gently prodded, respiratory distress at rest as indicated by deep abdominal excursions, or weight loss of ≥20%. Alveolar type 2 (AT2) cell isolation. AT2 cells were isolated as previously described 44 . Briefly, lungs were perfused with 0.9% saline, followed by inflation with 2. Western blot analysis. AT2 cells were harvested and lysed by sonication in 1X PBS containing 1% mammalian protease inhibitor cocktail (Sigma Aldrich) and 1X PhosSTOP (Sigma Aldrich) at a concentration of 1 × 10 6 cells/80 μl. Protein concentration in the supernatants was assessed with the Pierce Micro BCA Kit (Thermo Fisher Scientific 23235). Equal amounts of proteins were separated on 10-20% tris-tricine gels (EC66252BOX, Thermo Fisher Scientific), or 10-20% tris-glycine gels (XP10202BOX, Thermo Fisher Scientific) under reducing conditions at 125 V for 1.5 hours and transferred to 0.1 μm nitrocellulose membranes (GE Amersham 10600000) or 0.2 μm Immobilon-PSQ PVDF membranes (EMD Millipore ISEQ. 10100) at 180 mA for 1 hour using a semi-dry apparatus. Membranes were blocked in 5% non-fat dry milk and subsequently incubated with primary antibodies (Supplementary Table S3) overnight at 4 °C. Membranes were incubated with appropriate HRP conjugated secondary antibodies (Supplementary Table S4), developed using Immobilon forte Western HRP substrate (EMD Millipore WBLUF0100) and analyzed on ChemiDoc Touch Imaging System (BioRad). Membranes were stripped with Restore Western blot stripping buffer (Thermo Fisher Scientific). Densitometric analysis was performed using ImageLab software (BioRad). Histology, immunofluorescence and TUNEL assay. Lungs were inflated with 4% paraformaldehyde under 25 cm pressure and immersed in the same fixative overnight at 4 °C. Right lung lobes and left lung were sub-dissected and embedded in paraffin after dehydration in an ethanol series. Lung pieces were sectioned at 5 μm using a Leica RM2235 microtome for hematoxylin and eosin (H&E) staining. Tile scans of H&E stained sections were obtained using a Nikon NiE upright microscope. For immunofluorescence analyses, sections were stained with primary antibodies (Supplementary Table S3) with standard sodium citrate antigen retrieval as required. All Alexa-fluor conjugated secondary antibodies (Supplementary Table S4) were used at a dilution of 1:200. TUNEL assay (Roche 11684817910) was performed according to manufacturer's instructions on immunofluorescent labelled lung sections. High magnification images were taken at 60X with Nyquist magnification using Nikon A1 LUNA inverted microscope or at 100X using Nikon A1R LUNV inverted microscope. Confocal images in Fig. 6 were deconvoluted using Lucy-Richardson algorithm with 15 iterations on Nikon Elements. Immune cells were isolated from BALF by centrifugation at 230 g for 10 minutes, and pellets were resuspended in 100 μl of 0.9% saline for total and differential cell analyses. Cell-free BALF was used for measurement of surfactant (2019) 9:12509 | https://doi.org/10.1038/s41598-019-49020-4 www.nature.com/scientificreports www.nature.com/scientificreports/ lipids and total protein. Lipids were extracted from cell-free BALF by the Bligh and Dyer method 50 . Saturated phosphatidylcholine was isolated using the osmium tetroxide based method of Mason et al. 51 and quantitated by phosphorous measurement. Cytospin slides were prepared at a cell density of 1 × 10 6 cells/ml and stained with the Shandon Kwik-Diff stain kit (Thermo Fisher Scientific). A total of 300 cells was counted manually per slide to determine the percentage of monocyte/macrophages, lymphocytes and neutrophils. transmission electron microscopy. Mouse lungs were inflation-fixed with 2% paraformaldehyde [Electron Microscopy Sciences (EMS)], 2% glutaraldehyde (EMS), 0.1% calcium chloride (Fisher Scientific) in 0.1 M sodium cacodylate (EMS) buffer (SCB), pH 7.2, under 25 cm pressure, followed by immersion fixation with fresh fixative at 4 °C overnight. Lung lobes were cut into 1-2 mm blocks and processed for transmission electron microscopy as previously described 52 . 90 nm mouse lung sections were viewed, and images were digitally acquired by a Hitachi H-7650 transmission electron microscope (Hitachi High Technologies USA) equipped with a CCD camera (Advanced Microscopy Techniques) at 80 kV.
Scanning electron microscopy. Paraffin embedded mouse lung blocks used for confocal microscopy were dewaxed, rehydrated, and processed for scanning electron microscopy. Dewaxed mouse lung blocks were washed with HemoDe ® (EMS) to remove excess paraffin, followed by rehydration through a graded series of alcohol and distilled water mixtures 53 . To enhance specimen contrast and surface conductivity, rehydrated mouse lung blocks were first incubated with 1% osmium tetraoxide (OsO 4 ; EMS) and 1.5% potassium ferrocyanide (EMS) in 0.1 M SCB, pH 7.2, at room temperature for 2 hours, rinsed with 0.1 M SCB, incubated with freshly prepared 1% thiocarbohydrazide (Sigma) for 30 minutes, and fresh 1% OsO 4 in 0.1 M SCB for an additional hour. After osmium impregnation, en bloc staining with 4% uranyl acetate (EMS) was conducted at 4 °C overnight, followed by freshly prepared Walton's lead aspartate staining en bloc at 60 °C for 1 hour 54,55 . Lung blocks were dehydrated through a series of graded alcohol and processed for critical drying using a Leica EM critical point dryer (Leica Microsystems Inc.). Mouse lung samples were mounted on 15 mm specimen stubs (EMS) and viewed with a Hitachi SU-8010 field emission scanning electron microscope (Hitachi High Technologies USA) at 5 kV without further coating.
Barrier permeability measurements. Epithelial barrier permeability was measured as previously described 56 . Briefly, mice were intravenously injected with 300 μl of 12 mg/ml albumin conjugated to FITC (Sigma-Aldrich A9771), under 2% isoflurane anesthesia. BALF was recovered after 2 hours by lavage with 1 ml of 0.9% saline, and whole blood was collected via cardiac puncture. FITC fluorescence in BALF and serum was determined using the Synergy2 Multimode Microplate Reader (BioTek) with absorption/emission wavelengths of 480/520 nm. Barrier permeability was defined as the ratio of fluorescence in BALF to fluorescence in serum.
RNA sequencing and analysis. Total RNA was extracted from the entire pool of freshly isolated AT2 cells using RNeasy mini prep kit (Qiagen) and sent to Novogene Corporation (Chula Vista, CA) for RNA sequencing. mRNA was purified from total RNA (RNA integrity number >9) and sequencing libraries were generated using NEBNext ® Ultra ™ RNA library prep kit for Illumina. Sequencing was carried out on Illumina HiSeq 4000 platform and paired-end reads were generated. Filtered reads were aligned to the mouse reference genome mm10 using STAR (version 2.5). Differential expressions were determined through DESeq2 R package (version 2_1.6.3). Resulting p-values were adjusted using Benjamini and Hochberg's method for controlling false discovery rate. Genes with an adjusted p-value < 0.05 were assigned as differentially expressed. Gene ontology (GO) analysis was performed using Toppfun on Toppgene suite (https://toppgene.cchmc.org/) with Bonferroni method for multiple correction and 0.05 cut-off for significance, or clusterProfiler R package (version 2.4.3) with 0.05 cut-off for adjusted p-value. Heatmaps were generated using Heatmapper (http://www2.heatmapper.ca/expression/). Native gel analysis and proteasome activity assay. Native gel electrophoresis and proteasome activity assay were performed as previously described 14,57 . Briefly, isolated AT2 cells were lysed with ice cold lysis buffer (50 mM Tris-HCl, pH 7.5, 0.5% NP-40, 1 mM DTT, 2 mM ATP, and 5 mM MgCl 2 ) at a concentration of approximately 1 × 10 6 cells/100 μl by repeated pipetting. Protein concentration in the cell lysates was determined using the micro BCA kit. Equal amounts of protein were separated on 3-8% tris acetate gels (EA0378BOX, Life Technologies) at 150 V for 4 hours at 4 °C. Gels were incubated for 30 minutes at 37 °C in assay buffer (50 mM Tris, pH 7.5, 10 mM MgCl 2 , 1 mM ATP, 1 mM DTT) containing 50 μM Suc-LLVY-AMC (Sigma S6510) to monitor chymotrypsin-like activity of the 26S proteasome. Gels were imaged on Chemidoc touch imaging system with a UV transilluminator (AMC excitation/emission wavelengths are 380/460 nm). Subsequently, SDS was added to a final concentration of 0.02% and gels were incubated in assay buffer for an additional 20 minutes at 37 °C to image the latent activity of the 20S catalytic core. Gels were washed briefly in 50 mM Tris, pH 7.5 and incubated in 2% SDS, 66 mM Na 2 CO 3 , 1.5% β-mercaptoethanol for 10 minutes prior to semi-dry transfer to Immobilon-PSQ PVDF membranes at 250 mA for 1.5 hours. Proteasome complexes were immunoblotted as described in the Western blotting section. Imperial protein stain (Thermo Fisher Scientific 24615) was performed to assess protein loading. Human 26S proteasome (Enzo Life Sciences BML-PW9310-0050) and bovine 20S proteasome (UBPBio A1400) were used as controls.
Statistics. Statistical analysis was performed using GraphPad Prism (GraphPad Software). All data are presented as mean ± SD. Statistical tests used for comparison of data are reported in the respective figure legends.

Data Availability
RNA sequencing data is available on GEO (GSE123873). List of all differentially expressed genes and GO terms used to generate heatmaps are provided in supplementary file 2.