Autophagy, TERT, and mitochondrial dysfunction in hyperoxia

Abstract

Ventilation with gases containing enhanced fractions of oxygen is the cornerstone of therapy for patients with hypoxia and acute respiratory distress syndrome. Yet, hyperoxia treatment increases free reactive oxygen species (ROS)-induced lung injury, which is reported to disrupt autophagy/mitophagy. Altered extranuclear activity of the catalytic subunit of telomerase, telomerase reverse transcriptase (TERT), plays a protective role in ROS injury and autophagy in the systemic and coronary endothelium. We investigated interactions between autophagy/mitophagy and TERT that contribute to mitochondrial dysfunction and pulmonary injury in cultured rat lung microvascular endothelial cells (RLMVECs) exposed in vitro, and rat lungs exposed in vivo to hyperoxia for 48 h. Hyperoxia-induced mitochondrial damage in rat lungs [TOMM20, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT)], which was paralleled by increased markers of inflammation [myeloperoxidase (MPO), IL-1β, TLR9], impaired autophagy signaling (Beclin-1, LC3B-II/1, and p62), and decreased the expression of TERT. Mitochondrial-specific autophagy (mitophagy) was not altered, as hyperoxia increased expression of Pink1 but not Parkin. Hyperoxia-induced mitochondrial damage (TOMM20) was more pronounced in rats that lack the catalytic subunit of TERT and resulted in a reduction in cellular proliferation rather than cell death in RLMVECs. Activation of TERT or autophagy individually offset mitochondrial damage (MTT). Combined activation/inhibition failed to alleviate hyperoxic-induced mitochondrial damage in vitro, whereas activation of autophagy in vivo decreased mitochondrial damage (MTT) in both wild type (WT) and rats lacking TERT. Functionally, activation of either TERT or autophagy preserved transendothelial membrane resistance. Altogether, these observations show that activation of autophagy/mitophagy and/or TERT mitigate loss of mitochondrial function and barrier integrity in hyperoxia.

NEW & NOTEWORTHY In cultured pulmonary artery endothelial cells and in lungs exposed in vivo to hyperoxia, autophagy is activated, but clearance of autophagosomes is impaired in a manner that suggests cross talk between TERT and autophagy. Stimulation of autophagy prevents hyperoxia-induced decreases in mitochondrial metabolism and sustains monolayer resistance. Hyperoxia increases mitochondrial outer membrane (TOMM20) protein, decreases mitochondrial function, and reduces cellular proliferation without increasing cell death.

INTRODUCTION

Ventilation with gases containing enhanced fractions of oxygen is a cornerstone of therapy for patients with hypoxia and acute respiratory distress syndrome (ARDS), but is injurious in and of itself (1). Patients with COVID-19 pneumonia develop ARDS and require periods of extended exposure to high fractions of oxygen (2, 3), more than twofold longer than that of patients with other viral pneumonias. Thus, the need to better understand mechanisms of oxygen toxicity to lung tissue is abundantly clear.

We previously reported that exposure of cultured pulmonary artery endothelial cells to hyperoxia for 48 h enhances mitochondrial fragmentation via increased expression and phosphorylation of Drp-1, decreased Mfn-1, and increased expression of p62, Pink1, and LC3B. These responses to hyperoxia were prevented and rescued by initiation of mitochondrial DNA repair with mitochondrial targeted endonuclease III (4). The findings are also consistent with our previous observations in rats exposed in vivo to hyperoxia, demonstrating decreased mitochondrial complex I and II activities in lung homogenates (by 77% and 63%, respectively) relative to rat lungs exposed to normoxia, consistent with a more reduced mitochondrial matrix in hyperoxic lungs than their normoxia counterparts (5). The role of autophagy and mitophagy in cellular stress responses is well established (6) but has been less extensively studied in in vivo models of pulmonary injury by hyperoxia. Whether autophagy is activated because of mitochondrial dysfunction (e.g., fission activates autophagy) or whether mitochondrial dysfunction impairs autophagy in hyperoxic lung injury is unknown. Also, there is very little information regarding hyperoxic changes in mitochondrial density or mitochondrial mutations in lungs. Studies in the present work address these questions and were undertaken with the hypothesis that activation of autophagy and mitophagy would mitigate hyperoxic lung injury.

The extranuclear contributions of the catalytic subunit of telomerase, telomerase reverse transcriptase (TERT), to cellular and mitochondrial health are emerging areas of research in distinction to its well-established canonical role in elongating chromosomal ends. Existing evidence supports a definite role for TERT as one of the governing factors that regulates both mitochondrial integrity and autophagy. Our previous work demonstrated that loss of TERT plays a critical role in the development of the microvascular phenotype associated with coronary artery disease, and upregulation of TERT activity can restore physiological levels of nitric oxide through the activation of autophagy (7, 8). Similarly, loss of TERT catalytic function augments adverse cardiovascular phenotypes (9), whereas increased expression or activity appears to be protective (10). The underlying molecular nature of the protective effect is not clearly defined, but TERT can bind to mitochondrial DNA (_mtDNA) and may protect it from oxidative damage (11, 12), providing a potential mechanism by which _mtTERT can protect from hyperoxia-induced _mtDNA damage. Using a TERT loss-of-function rat model (9) and cultured lung microvascular endothelial cells, with pharmacological regulators of autophagy and mitophagy, we investigated the role of TERT in the response of rat lungs to hyperoxia, and in particular the role of TERT, as it impacts autophagy and/or changes in mitochondrial function during hyperoxia exposure. Our working hypothesis was that, like autophagy/mitophagy, TERT expression protects from mitochondrial damage in hyperoxia.

MATERIALS AND METHODS

Reagents

Antibodies against Pink1, TOMM20, and myeloperoxidase (MPO) were purchased from Abcam Incorporation (ab56783, ab186735, and ab188211, respectively). Antibodies to LC3B, Parkin, and Beclin-1 were obtained from Cell Signaling Technology (Danvers, MA; 3455, 2132, and 3738, respectively). Antibody to p62 was purchased from Abnova (H00008878-M01). Antibody to IL-1β was purchased from R&D (MAB5011). TERT antibody was obtained from Bioss (bs-1411R). The HEK-Blue hTLR4 and hTLR9 cells were obtained from InvivoGen (hkb-htlr4 and hkb-htlr9, respectively). The OxPhos Rodent WB Antibody Cocktail was purchased from Invitrogen (45–8099). β-Actin antibody (A2228), bafilomycin A1 (BAFA1, final concentration 100 nmol/L), and trichostatin A (TSA, final concentration 100 nmol/L) were purchased from Sigma-Aldrich. Other reagents purchased were trehalose (Research Products International, final concentration 10 mmol/L) and BIBR 1532 (Tocris Bioscience; final concentration 10 µmol/L). AGS 499 was a gift from Ester Priel, PhD, Israel (final concentration 250 nmol/L).

In Vitro Hyperoxia Exposure of Cells

Rat lung microvascular endothelial cells (RLMVECs) were purchased from Cell Biologics (M1266). Cells between passages 3 and 5 were used for the studies in this work. After lifting, cells were plated at a density such that they were 70% confluent in 1–3 days in standard cell culture conditions (humidified air, 5% CO₂, 37°C). Once they reached ∼70% confluence, vehicle or test agents were added to the culture media, and plates were transferred to a hyperoxic environment (95% ${\text{FI}}_{{\text{O}}_2}$, 5% CO₂) or left in a normoxic cell culture environment with room air supplemented with 5% CO₂ for 48 h. A normoxic oxygen percentage of 21% was chosen for in vitro studies because lung microvascular endothelial cells function in approximately this partial pressure of oxygen. Cultured cell experiments were performed with at least three biological and three technical replicates each (4).

TERT Knockout Rats

The rat knockout for TERT, the catalytic subunit of telomerase, was developed on the outbred Sprague–Dawley (SD) background using CRISPR/Cas9 technology, as previously described (9). A CRISPR guide RNA targeting the first exon sequence GGGCAACGAGGAGCGCGGGG of TERT (protospacer adjacent motif) was successfully used to generate a 17-bp frame shift mutation following pronuclear injection into SD/NCrl rat embryos, resulting in a premature stop codon and effectively eliminating functional TERT expression. The genetic modification is in the first part of the coding region and can change regulation of expression if a nonspecific frame shift mutation has occurred; regulation of expression in both strains is not known to be different. Tissue from the TERT knockout (TERT KO) rats exhibit essentially no functional TERT activity in telomeric repeat amplification (TRAP) assay (9), despite exhibiting immune-specific protein to TERT antibodies. A heterozygous breeding colony of SD-TERT animals was established; homozygous TERT mutant rats are viable and born at Mendelian frequencies. For all studies, WT littermates were used as controls for TERT KO in normoxia and hyperoxia studies.

In Vivo Hyperoxia Exposure

Treatment protocols were approved by the Institutional Animal Care and Use Committees of the Zablocki Veterans Affairs Medical Center, the Medical College of Wisconsin, and Marquette University (Milwaukee, WI). Heterozygous rats were excluded from the study and no rats were excluded post hoc after exposures/treatments were begun. For experiments with WT and KO rats, littermates were exposed in parallel whenever available. For normoxia (control) rat studies, male and female rats (WT: 370 ± 64 g, n = 12; TERT KO; 324 ± 44 g, n = 13) were exposed to room air in plexiglass chambers side by side, with those exposed to hyperoxia (WT: 311 ± 70 g, n = 46; TERT KO; 339 ± 50 g, n = 40). Rat ages were between 54 and 112 days, with a median of 76 days. No phenotypic changes were observed between sexes or age, so results are reported together (13, 14).

In a subset of rats, 2% trehalose was added to the drinking water to activate autophagy for 28 days, as previously published, before exposure to hyperoxia or normoxia for 48 h, during which time the trehalose or vehicle was continued (15, 16).

Pulmonary Histology

Upon euthanasia, lungs were fixed with 10% neutral buffered formalin in the inflated state. Thick slices (4 µm) of paraffin-embedded, fixed whole mount sections of lung were stained for hematoxylin and eosin (H&E) or immunostained for CD68 (Millipore, MAB1435). High-resolution JPEG images of the whole mount sections were used for quantification of injury, according to methods previously published (17) or for macrophage cell counting using ImageJ.

To score injury, we used a modification of a scale that grades epithelial injury, fibrosis, and inflammatory changes (EFI), each on a zero to three scale (Table 1). Six representative images from each rat were acquired by one investigator then graded by a second investigator blinded to the treatment group.

Table 1.

Modified graded EFI scale

Scale	0	1	2	3
Epithelial	Normal	Reactive epithelium ± intermittent disruption of epithelial layer	Markedly reactive epithelium ± epithelial disruption	Epithelium largely denuded
Fibrosis	Normal	Scattered young fibroblasts	Intraluminal fibromyxoid filling	Dense fibrosis
Inflammation	Normal	Perivascular or peribronchiolar cuffing with mixed sparse inflammatory cells/sparse chronic inflammation	Moderate mixed inflammatory infiltration throughout the lungs, moderate edema	Severe infiltration of mixed inflammatory cells, edema, and thickened alveolar capillary membrane

EFI, epithelial injury, fibrosis, and inflammatory changes.

We quantified CD68 staining as a marker of tissue macrophages in lung sections from WT and KO rats exposed to normoxia and hyperoxia using ImageJ. The colors from the immunohistochemistry (IHC) images were deconvoluted. The blue hematoxylin image was thresholded and watershed to give the number of cell nuclei in the image, the brown DAB image was thresholded to give the number of macrophage-stained cells in the image, and the percent macrophages were calculated comparing the number of macrophages to the total cell counts. Six lungs per condition and 10 images at ×20 per lung were analyzed.

Western Blots for Cell and Lung Homogenates

Cultured cells were harvested by scraping in the presence of 0.5 mL of RIPA buffer (20–188; Millipore, Temecula, CA) supplemented with protease inhibitor cocktail (539134, Millipore) (18). Lung tissues dissected free of large airways or blood vessels were homogenized in the same buffers with protease inhibitors. The mixture was kept on ice for 15 min, after which lysates were centrifuged for 10 min at 20,000 g and the supernatants used for determining protein concentration via the Bio-Rad protein assay kit. Equal amounts of protein (10–50 μg/lane) were boiled for 5 min in Laemmli sample buffer (161–0737, Bio-Rad) supplemented with 2-mercaptoethanol, resolved on Tris·HCl SDS polyacrylamide gels (Bio-Rad), and transferred to nitrocellulose membranes. The blots were developed with specific antibodies against TERT (1:1,000), p62 (1:1,000), Pink1 (1:500), Parkin (1:1,000), LC3B (1:1,000), TOMM20 (1:5,000), OxPhos (1:2,000), and matched secondary antibodies and visualized using ECL Plus detection reagent (Thomas Scientific, 273988). Immune-specific protein levels were normalized to β-actin density except for the ratio of LC3B-II to I, where only LC3B was normalized. Images were scanned and the relative densities were determined by ImageJ software. All lanes and individual data points represent biological rather than technical replicates (13).

Mitochondrial Function/Cell Survival

3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay (19) was used as an indicator of mitochondrial function. In this colorimetric assay, the amount of purple formazan dye formed from MTT is proportional to the number of metabolically active cells and their mitochondrial activity. MTT assays were performed with either RLMVECs plated at a density of 5 × 10³ cells/well in a 96-well dish or lung pieces from WT or TERT KO rats. When cells were 70%–80% confluent, plates were subjected to different experimental treatments, as described in the text (normoxia vs. hyperoxia with TERT and autophagy inhibitors or activators). After 48 h, media was removed, and the cells were incubated for 3 h in phenol red-free medium containing 0.5 mg/mL MTT. After 3 h, media was removed and 200 mL DMSO added to each well. After 10 min, absorbance at 540 nm was measured in an ELISA plate reader (PCR ELISA plus Roche, 12013789001).

For lung pieces, 40–50 mg tissue were placed in 0.6 mL MTT solution (DPBS buffer with 0.5 mg/mL MTT) at 37°C for 2 h. At the end of this time, 1 mL DMSO was added, the lung pieces homogenized, and the homogenate incubated at room temperature for 2 h. Lung homogenates were then centrifuged to remove cellular debris and the absorbance in the supernatant read at 540 nm in an ELISA plate reader.

Propidium Iodide and Hoechst Assays

Propidium iodide (PI; Sigma, K4170) and Hoechst 33342 (Thermo Fisher Scientific, 62249) assays on RLMVECs were performed in 96-well plates, as per manufacturer’s instructions, and read on a SpectraMax M5 (Molecular Devices) plate reader. Unbound PI emission and excitation wavelengths were 493/636, bound PI emission and excitation wavelengths were 535/617, and Hoechst emission and excitation wavelengths were 361/497. Cell membrane integrity excludes PI from staining viable and apoptotic cells, and thus PI identifies dead cells. Hoechst 33342 is cell permeable and binds to DNA, and thus detects both living and apoptotic cells. When these dyes are used in combination in the same sample and at the same time, intact cells, cells undergoing apoptosis, and dead cells resulting from apoptotic and/or necrotic processes can all be identified.

Mitochondrial/Nuclear DNA Ratio and Mitochondrial DNA Damage

_mtDNA damage was quantified using quantitative PCR-based methods (20), as previously reported. In brief, total DNA was extracted using the QIAamp DNA Blood Mini Kit (QIAGEN, Valencia, CA). Primers specific to the mitochondrial genome (Dloop, ATPase 6/8, ND 1/2, and ND 4/5 regions) that result in a short product and a long product for each site were used for amplification. The short and the long products were compared with each other to determine the lesions per 10 kb at each site. Mitochondrial/nuclear DNA (_mtDNA/_nucDNA) ratio was determined by comparing the amplification of the short fragments from the mitochondrial PCRs to the amplification of a short fragment of a single-copy nuclear gene (Fars2). Relative amplifications were calculated to compare WT and TERT KO lungs after normoxia or hyperoxia treatment.

TLR4 and TLR9 Plasma Assays

TLR4 and TLR9 activation by components within the rat plasma were evaluated using the HEK-Blue-hTLR4 and HEK-Blue-hTLR9 cells (InvivoGen). The cells were grown and used according to the manufacturer’s protocol. In brief, plasma and a control standard (LPS for TLR4 or ODN2006 for TLR9) were plated in 96-well dishes with the HEK-Blue cells overnight. If the plasma activates either TLR4 or TLR9, an inducible secreted embryonic alkaline phosphatase reporter is activated, turning the media blue, which is quantified with a spectrophotometer at 620 nm.

Transwell Resistance Assays

Cells were plated at a density of 1 × 10⁴ cells per well in a transwell with 0.4 μm pores (Corning, 3470) set inside a 24-well culture dish. A volume of 500 μL media was used in the transwell and 750 μL in the outside well. Monolayer resistance was measured daily after 48 h using a WPI Epithelial Volt/Ohm [transendothelial electrical resistance (TEER)] Meter. When monolayer resistance reached a steady state, an activator of telomerase transcription (AGS 499), autophagy activators (TSA or trehalose), or vehicle were added to the wells then the plates were placed in normoxia or hyperoxia for 48 h. After 48 h, monolayer resistance was measured and data normalized to that of concomitant vehicle controls (4, 21).

Wet-to-Dry Weight Lung Measurements

Heart and lungs from a randomly selected subset of each of the groups of rats were isolated and washed free of blood using the isolated perfused lung preparation described below for filtration coefficient measurements (K_f). After washing, the lungs were dissected free of the heart, trachea, and mainstem bronchi and total lung wet weight was obtained. The left lung lobe was weighed and dried at 60°C until dry weight was stable for 2 days for wet-to-dry weight ratio (14).

K_f Measurements

K_f, a measure of vascular permeability, was determined using the approach previously described by us. Representative rats from normoxia or 48 h hyperoxia were anesthetized with Beuthanasia (40–50 mg/kg ip). The trachea was surgically isolated and cannulated, the chest opened, and heparin (0.7 IU/g body wt) injected into the right ventricle. The pulmonary artery and the pulmonary venous outflow were accessed via cannula. The lung and heart were removed and suspended from a calibrated force displacement transducer (ModelFT03; Grass Instruments) and weight was monitored continuously. The Krebs-Ringer bicarbonate perfusate contained (in mM) 4.7 KCl, 2.51 CaCl₂, 1.19 MgSO₄, 2.5 KH₂PO₄, 118 NaCl, 25 NaHCO₃, 5.5 glucose, and 5% bovine serum albumin. The perfusion system was primed (Masterflex roller pump) with the perfusate maintained at 37°C and equilibrated with 15% O₂-6% CO₂-balance N₂ gas mixture resulting in perfusate Po₂ (∼105 Torr), Pco₂ (40 Torr) and pH 7.4. The lung was ventilated (40 breaths/min) with the above gas mixture and end-inspiratory and end-expiratory pressures of ∼6 and 3 mmHg, respectively. The pulmonary arterial and venous pressures were referenced to atmospheric pressure at the level of the left atrium and monitored continuously during the course of the experiments. Perfusate was pumped (0.03 mL/min/g body wt) through the lung until it was evenly blanched and venous effluent was clear of visible blood before switching from single pass to recirculation mode (13).

After a 10-min stabilization period with the venous pressure (P_V) set at atmospheric pressure, P_V was raised to 3.7 mmHg and the lung perfused for 20 min. Then P_V was raised to 10 mmHg and perfused for an additional 10 min. K_f was determined by dividing the difference in the rate of lung weight gain measured 10 min after increasing P_V from 3.7 to 10 mmHg and after increasing P_V from 0 to 3.7 mmHg by the difference in pulmonary capillary pressure at these P_V values. For each P_V, the capillary pressure was estimated as the averages of arterial pressure and venous pressure. K_f was normalized to gram of dry lung weight.

Statistical Analysis

Data are presented as means ± SE. Statistical analysis was performed with one-way ANOVA followed by Tukey’s or Shapiro–Wilk tests where appropriate. When comparing only two conditions (normoxia vs. hyperoxia) or the response of a treatment versus vehicle, t tests were employed where indicated in the text or figure legends. P ≤ 0.05 was considered statistically significant. All experimental n’s appear in the figure legends or figures.

RESULTS

Hyperoxia Increases Pulmonary Histological Injury

To assess pulmonary injury, we acquired H&E images from fixed inflated lungs from WT or KO rats exposed in vivo for 48 h to hyperoxia or normoxia (Fig. 1, A–D). Images were graded for epithelial, inflammatory, and fibrotic changes. There were no differences between the epithelial injury scores of WT and KO rats in normoxia or hyperoxia (Fig. 1I) and no fibrotic changes in any samples (data not shown). Both WT and KO lungs samples exhibited an increase in inflammatory changes in response to hyperoxia, but these responses were not different between the two strains (Fig. 1J). Due to the increased inflammation, we looked at levels of tissue macrophages by immunostaining the IHC slides with a broad macrophage marker, CD68 (Fig. 1, E–H). No differences in macrophage levels between WT and KO rats in normoxia or hyperoxia were observed (Fig. 1K).

Figure 1.

Hyperoxia causes changes in inflammation. Representative histological images stained with hematoxylin and eosin (H&E) of lungs of wild-type (WT) (A and C) or TERT knockout (KO) (B and D) rats exposed in vivo to normoxia (A and B) or hyperoxia (C and D) for 48 h. The bar is 100 µm in length. The insets are enlarged to show inflammatory cells in hyperoxic samples. Green arrow depicts reactive epithelium and the yellow arrow indicates edema, consistent with hyperoxic lung injury. Representative images of lungs of WT (E and G) or KO rats (F and H) exposed in vivo to normoxia or hyperoxia then stained for macrophages with CD68. The bar is 100 µm in length. The percentage of CD68 positive cells appears to be similar in both conditions for WT and KO strains of rats. H&E stained images from WT and TERT KO rats exposed to hyperoxia or normoxia for 48 h (n = 6–8 rats in each group) were graded for epithelial and inflammatory changes (see Table 1). There were no differences in epithelial injury scores in any of the four groups (I), nor were there any fibrotic changes in any lung sections (scores were zero in all samples; data not shown). WT and TERT KO rats exhibited increased inflammation in hyperoxia relative to normoxia (J). K: percentage of CD68 positive cells/condition (n = 4 rat lungs each condition), which was not changed by hyperoxia in either strain. *P < 0.05, t test hyperoxia vs. normoxia.

Because we identified histological evidence of inflammation after hyperoxia, we next tested for MPO and mature IL-1β expression in lung homogenates. Figure 2, A and B shows an increase in MPO light chain in rats exposed to hyperoxia, both in WT and KO animals. However, MPO light chain in normoxic KO rats was increased over that of WT counterparts. IL-1β is a cytokine produced by activated macrophages that is proteolytically processed to its active form by caspase 1 (CASP1/ICE) and serves as an important mediator of the inflammatory response. We observed an increase in mature IL-1β levels in WT and KO rat lungs after hyperoxia (Fig. 3, A and B).

Figure 2.

Myeloperoxidase (MPO) in lungs of wild-type (WT) and knockout (KO) hyperoxic rats. Representative (A) and averaged (B) data from Western blots to assess expression of MPO light chain in the lungs of rats exposed in vivo to hyperoxia. In normoxic conditions, there is an increase in MPO in KO relative to WT rats. MPO increases substantially in tissue from both WT and KO rats with hyperoxia (n = 4 rat lungs per each condition). *P < 0.05, t test hyperoxia vs. normoxia and WT vs. KO in normoxia.

Figure 3.

Expression of IL-1β increases in lungs of wild-type (WT) and knockout (KO) hyperoxic rats. Representative (A) and averaged (B) data from Western blots to assess expression of mature IL-1β in the lungs of rats exposed in vivo to hyperoxia. Density of IL-1β is increased by hyperoxia in both strains of rats (n = 4–6 rat lungs per each condition). *P < 0.05, t test hyperoxia vs. normoxia. Because the same blots were stripped after probing for MPO and then reprobed for IL-1β, β-actin blots are the same for these Western blots (Figs. 2 and 3).

TERT Expression Decreases in Response to Hyperoxia Exposure

TERT is found in several naturally occurring isoforms (22–24). The primary antibody we used for Fig. 4 detects multiple isoforms visible in samples from cultured cells (see Supplemental Material for full Western blots including those with an additional antibody). In this work, we focused on the full-length TERT isoform found at 127 kDa. Full-length TERT protein expression in RLMVECs (Fig. 4A) and in lungs from rats (Fig. 4B) exposed to hyperoxia was decreased relative to that of normoxia samples.

Figure 4.

Levels of TERT decrease in response to hyperoxia exposure in wild-type (WT) rats. Representative and averaged data from Western blots to assess expression of the catalytic subunit of telomerase TERT are shown. Expression of TERT was modestly reduced in rat pulmonary microvascular endothelial cells (A) [rat lung microvascular endothelial cells (RLMVEC), n ≥ 4 biological replicates] and lungs from WT and knockout (KO) rats (B) (n ≥ 6 rat lungs in each group) exposed to 48 h of hyperoxia vs. normoxia. Protein levels were normalized to density of β-actin and expressed as fold change to normoxia WT. *P < 0.05, t test vs. control.

Hyperoxia Affects _mtDNA/_nucDNA Ratios

In cultured RLMVECs, 48 h of hyperoxia increased _mtDNA/_nucDNA ratios compared with that of normoxic controls (Fig. 5A). However, hyperoxia did not change _mtDNA lesions/10 kb in the Dloop, ND1/2, ND4,5, or ATPase6/8 regions of the mitochondria (Fig. 5B). In contrast to cells, hyperoxia exposure decreased the ratio of _mtDNA/_nucDNA in lungs of WT and KO rats (Fig. 5, C and E). Similar to cells, mitochondrial lesions/10 kB were not different in lungs of WT or KO rats exposed to hyperoxia (Fig. 5, D and F).

Figure 5.

Hyperoxia, mitogenesis, and _mtDNA damage. 48 h of hyperoxia exposure increased _mtDNA/_nucDNA ratios in cultured rat microvascular endothelial cells compared with that of normoxic controls (A) but did not cause changes in mtDNA damage (lesions/10 kb, B). In lungs of wild-type (WT) (C and D) and knockout (KO) rats (E and F) after in vivo exposure to hyperoxia, we observed decreased _mtDNA/_nucDNA ratios and no increase in mitochondrial lesions/10 kb. *P < 0.05, t test vs. normoxic control. For rat microvascular endothelial cells, n = 3 biological replicates and 3 technical replicates per biological sample. For rat lungs, n ≥ 4 per group.

Hyperoxia Increases TLR4 and TLR9 Activation

Because we saw modified _mtDNA/_nucDNA ratio with hyperoxia an increase in lung inflammation, we assayed rat plasma for two common damage associated molecular patterns (DAMPs): mitochondria membrane fractions recognized by TLR4 receptors and free mitochondrial DNA, which is recognized by TLR9. Hyperoxia caused an insignificant (P = 0.06; Fig. 6A) increase in plasma TLR4 and a significant increase in TLR9 after hyperoxia in WT but not TERT KO rats (Fig. 6B). Increases in TLR9 activating substances in hyperoxia support IHC data showing increased inflammation during hyperoxia.

Figure 6.

Plasma levels of TLR4- and TLR9-activating substances in hyperoxia. Plasma levels of TLR4- (A) and TLR9- (B) activating fragments were assessed using HEK-blue cells overexpressing receptors for either TLR4 or TLR9 receptors. Hyperoxia increased plasma TLR9-activating fragments in WT but not KO rats. There was a trend for hyperoxia to increase in TLR4, which did not reach statistical significance. n = 4 for each group; *P < 0.05.

TOMM20 Expression Is Increased in Hyperoxia

We probed cell and lung homogenates with primary antibodies to recombinant TOMM20 as a marker of mitochondrial outer membrane protein expression. Hyperoxia increased TOMM20 in cultured RLMVECs (Fig. 7, A and B) and in lungs of WT and KO rats (Fig. 7, C–E). TOMM20 expression in hyperoxic KO lung was greater than that of WT hyperoxic lungs (Fig. 7F).

Figure 7.

Increased outer membrane mitochondrial protein in rat lung microvascular endothelial cells (RLMECs) and lungs of rats exposed to hyperoxia and with TERT knockout (KO). A: representative RLMVEC homogenates probed with a primary antibody to TOMM20 are shown. Samples were also probed for β-actin to normalize for protein loading. B: averaged data for RLMVEC normoxia and hyperoxia conditions are shown. n ≥ 6 for each condition. C–F: TOMM20 protein was increased in KO rats under normoxia relative to wild-type (WT) cohorts. Moreover, TOMM20 was higher in KO hyperoxia than WT hyperoxia rat lung and KO normoxia. n = 4–11, *P < 0.01.

Hyperoxia Modifies Expression of Mitochondrial Complexes

To test whether the increases in TOMM20 and the _mtDNA/_nucDNA ratio changes in response to hyperoxia are associated with changes in the mitochondrial complexes, we ran Western analyses of WT and KO lung homogenates using an antibody cocktail containing antibodies to each of the five mitochondrial complexes (Fig. 8A). These data show clear bands consistent with complexes I through V in all samples. There were no differences in complex expressions in WT and KO lungs under normoxia. Complex I (CI) expression was decreased by hyperoxia in both WT and KO rats. Complex II (CII) expression was decreased in WT rats exposed to hyperoxia, but not KO rats (Fig. 8, B and C). The expression of other oxidative phosphorylation proteins was not different between lungs from rats in hyperoxic conditions.

Figure 8.

Western blots of mitochondrial complex expression. A: representative Western blots probed with an OxPhos antibody cocktail are shown with samples from each of the four groups of rats studied. The bands are shown separated because longer exposures were required to quantify CI relative to the remainder of the bands. The densest band of all the samples was cardiovascular (CV), with the lightest band being CI. B: averaged data depicting relative densities of all complexes in wild-type (WT) rats under normoxic and hyperoxic conditions are shown. Expression of mitochondrial complexes I–V (n = 5–7 for each condition) were not different in normoxic conditions for WT and knockout (KO) rat tissues. C: averaged data depicting densities of KO rats under normoxic and hyperoxic conditions. When exposed to hyperoxia, the lungs of WT rats exhibited a decrease in CI and CII complex relative to their normoxic counterparts. A similar decrease in CI, but not CII was observed in lungs from KO rats exposed to hyperoxia. (*P < 0.01).

Hyperoxia Modifies Mitochondrial Function

To test whether changes in mitochondrial protein levels have functional consequences in our injury models, MTT (as an indirect marker of mitochondrial metabolism) was used. We observed decreased MTT in cultured RLMVEC after exposure to hyperoxia (Fig. 9). Telomerase activity inhibition (BIBR 1532) caused a similar magnitude of decrease in MTT as exposure to hyperoxia, while enhancing TERT transcription (AGS 499) partially prevented the hyperoxia-induced decrease in MTT (Fig. 9A). Inhibition of autophagy with BAFA1 reduced MTT in normoxia and hyperoxia. Activation of autophagy on the other hand, by either TSA or trehalose, reduced the effects of hyperoxia on MTT (Fig. 9B). To test the upstream or downstream nature of the protective effects of autophagy and TERT activation in hyperoxia, we inhibited one pathway while activating the other in hyperoxic RLMVECs. Autophagy activation via trehalose in the presence of BIBR 1532 to block TERT caused a reduction in MTT compared with vehicle in normoxia and did not protect against the effects of hyperoxia. Similarly, treatment with AGS 499 to increase telomerase levels in cells while inhibiting autophagy using BAFA1 led to a reduction in MTT in normoxic conditions and did not restore MTT activity in hyperoxic cells (Fig. 9C). These data suggest the importance of both pathways in parallel in protection from hyperoxic injury.

Figure 9.

Hyperoxia decreases MTT in a manner sensitive to telomerase expression and autophagy activation. A: decreased MTT was observed in cultured rat lung microvascular endothelial cells (RLMVECs) after exposure to 48 h of hyperoxia or after pharmacological modulation of telomerase activity with AGS 499 or BIBR 1532 in normoxia. Telomerase inhibition (BIBR 1532) in normoxia resulted in a similar decrease in MTT as exposure to hyperoxia alone; telomerase activation (AGS 499) partially prevented hyperoxia-induced decrease in MTT. B: inhibition of autophagy with bafilomycin (BAFA1) reduced MTT in normoxia, whereas activation of autophagy [trichostatin A (TSA) or trehalose] had no effect on mitochondrial function in normoxia. Activation of autophagy with TSA or trehalose partially mitigates hyperoxia-induced decrease in MTT, whereas inhibition of autophagy with bafilomycin further reduced mitochondrial function over hyperoxia alone but not in comparison with bafilomycin normoxia. C: autophagy activation with trehalose in the setting of telomerase inhibition (with BIBR 1532) is not sufficient to rescue hyperoxia-associated decrease in MTT. Similarly, stimulation of telomerase expression with AGS 499 in the setting of autophagy inhibition by BAFA1 is insufficient to compensate for hyperoxia-induced decrease in MTT. *P < 0.05, two-way ANOVA with Sidaks multiple comparisons vs. untreated normoxia control (t test). n = 5–17 biological replicates with 3+ technical replicates. These data suggest that both TERT activity and robust autophagy are required for optimal MTT activity. MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide.

Hyperoxia Does Not Induce Cell Death but Prevents Cellular Proliferation

We next investigated the effects of hyperoxia on cell death and proliferation to complement the MTT studies. As PI only enters dead or dying cells, we measured the ratio of bound to free PI after 48 h of normoxia or hyperoxia to determine whether hyperoxia was causing cell death in RLMVECs. Neither exposure to hyperoxia, treatment with autophagy modulators, nor treatment with TERT modulators changed the number of dead/dying cells (Fig. 10A). Hoechst is a cell-permeable DNA stain that binds preferentially to adenine-thymine (A-T) rich regions of DNA in live and dead cells. Therefore, this stain was used to quantify the total number of cells. Our data suggest that cells continue to proliferate during normoxia but hyperoxia impairs cell growth (Fig. 10B). In addition, the presence of inhibitors of autophagy or telomerase activity in either normoxia or hyperoxia led to a decrease in number of total cells (Fig. 10B). These data suggest that the results from the MTT measurements under these conditions reflect an effect on cell proliferation rather than cell death.

Figure 10.

Hyperoxia does not induce cell death but prevents cellular proliferation. A: rat lung microvascular endothelial cells (RLMVECs) were plated at ∼50% confluence and cultured with vehicle or inhibitors of autophagy or telomerase activity (BAFA1 or BIBR 1532, respectively) and placed in normoxia or hyperoxia at ∼70% confluence. Forty-eight hour later, dead cells were quantified using the ratio of bound vs. free propidium iodine (PI). Treatment with autophagy or telomerase activity modulators or exposure to 48 h of hyperoxia had no effect on the number of dead cells. n = 4–7 biological replicates with 5–8 technical replicates. B: Hoechst 33342 stain was used to quantify total number of cells. These data show a decrease in cells with inhibitors of autophagy (BAFA1) or telomerase activity (BIBR 1532) or both, as well as after exposure to 48 h of hyperoxia. Autophagy activation [trichostatin A (TSA)] or increased telomerase synthesis (AGS 499) had no effect on total cell numbers in normoxia or hyperoxia relative to vehicle control. n = 3–6 biological replicates with 5–8 technical replicates *P < 0.05, two-way ANOVA with Sidaks multiple comparisons vs. normoxia vehicle. #P < 0.05, two-way ANOVA with Sidaks multiple comparisons vs. hyperoxia vehicle. C: absence of the catalytic subunit of telomerase (TERT KO) decreased lung MTT in normoxia (baseline) relative to WT control. 48 h exposure to hyperoxia decreased mitochondrial metabolism in wild-type (WT) rats, but not in TERT knockout (KO) rats (which started lower than WT). Autophagy activation via treatment with trehalose partially restored MTT in normoxia or hyperoxia. *P < 0.05, two-way ANOVA with Sidaks multiple Comparisons vs. WT same treatment. #P < 0.05, two-way ANOVA with Sidaks multiple comparisons vs. KO same treatment. n = 5–16/group. MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide.

Autophagy Activation Counteracts Loss of TERT and Hyperoxia-Induced Decrease in Mitochondrial Metabolism In Vivo

Lung tissue from WT rats exhibited higher MTTs than that of TERT KO rats in normoxia (Fig. 10C). Autophagy activation by treatment with trehalose for 28 days rescued the decreased MTT in normoxic KO rats but had no effect on MTT of WT rats under this condition. WT rats exposed to hyperoxia showed decreased mitochondrial metabolism evaluated by MTT relative to that of tissue from normoxic WT rats, whereas there was no difference in MTTs from lungs of KO rats exposed to normoxia or hyperoxia. Autophagy activation by treatment with trehalose for 28 days partially rescued the decreased MTT after hyperoxia treatment in both WT and KO rats.

Loss of TERT Function Impairs Autophagosome Formation in Rat Lungs in Response to Hyperoxia

Hyperoxia caused a decrease in Beclin-1 in both WT and KO rat lung tissue (Fig. 11, A and B). Hyperoxia also resulted in a lower density of LC3B-II relative to LC3B-I and therefore a decreased ratio of LC3B-II/I in WT but not TERT KO rats (Fig. 11, C–F). Autolysosome clearance decreased in response to hyperoxia in both WT and KO, as indicated by elevated p62 levels with this exposure. No differences in p62 expression between WT and KO in either normoxia or hyperoxia (Fig. 11, G and H) were observed. Together these data are most consistent with hyperoxia-induced stalled autophagy and altered autophagic flux.

Figure 11.

Hyperoxia modifies Belcin-1 and cleavage of LC3B-I to LC3B-II in wild-type (WT) rat lungs. Representative Western (A) and averaged (B) data of Beclin-1 (protein size ∼60 kDa) in samples from lungs of WT and knockout (KO) rats exposed to normoxia and hyperoxia show that Beclin-1 expression decreases with hyperoxia in both WT and KO rats (n = 4 each group; *P < 0.05 t test vs. WT normoxia). C: representative Western of LC3B (protein size 19/15 kDa) in lungs of TERT WT and KO rats in normoxia and hyperoxia is shown. Hyperoxia decreases the density of LC3B-II in WT but not KO rats. Average relative LC3B-I densities (D), average relative LC3B-II densities (E), and LC3B-II/I ratios decrease in WT but not TERT KO rats exposed to hyperoxia (F). Representative Western (G) and averaged (H) data of p62 (protein size 62 kDa) in samples from lungs of WT and KO rats exposed to normoxia and hyperoxia show that p62 expression increases with hyperoxia in both WT and KO rats (n = 6 each group; *P< 0.05 t test vs. WT normoxia).

Mitophagy Is Not Altered by Loss of Telomerase Activity

Pink1 expression increased after hyperoxia exposure in both WT and TERT KO (Fig. 12, A and C) consistent with damaged mitochondria being flagged for delivery to the autophagosome. However, Parkin levels remained unchanged in both WT and KO rats exposed to hyperoxia stress (Fig. 12, B and D), suggesting no defect in the actual delivery of damaged mitochondria for recycling via autophagy during hyperoxia stress.

Figure 12.

Mitophagy is not changed by TERT. A: Pink1 (protein size 66 kDa) expression is elevated in response to hyperoxia in both wild-type (WT) and knockout (KO) rats with hyperoxia. B: Parkin (protein size 52 kDa) levels do not change in WT or TERT KO rats with hyperoxia. All protein densities were normalized to levels of β-actin (protein size 42 kDa) and expressed as fold change to WT normoxia control. *P < 0.05 t test vs. control n = 9–10. These data suggest that defective mitochondria in both TERT WT and KO rats are effectively being delivered for lysosomal degradation.

Increased Telomerase or Autophagy Activation Is Sufficient to Prevent Hyperoxia-Induced Loss of Endothelial Barrier Function

The consequences of telomerase activation or autophagy activation on endothelial barrier function were evaluated by transendothelial electrical resistance (TEER) measurements. Hyperoxia decreased barrier function in RLMVECs relative to normoxia samples (Fig. 13). Activation of telomerase reduced TEER under normoxic conditions, whereas both increased telomerase and initiation of autophagy with TSA or trehalose prevented hyperoxia-induced loss of TEER.

Figure 13.

Increased telomerase activity or autophagy activation is sufficient to prevent hyperoxia induced loss of endothelial barrier function. Endothelial barrier function was evaluated by transendothelial electrical resistance (TEER). Activation of telomerase reduced TEER under normoxic conditions, whereas hyperoxia decreased TEER. Either telomerase activation or initiation of autophagy prevented hyperoxia-induced loss of TEER. Autophagy activation had no effect on TEER in normoxia. *P < 0.05 test vs. vehicle; #P < 0.05 vs. vehicle hyperoxia; n = 3 biological replicates with 4+ technical replicates each. TSA, trichostatin A.

K_f, Wet-to-Dry Lung Weights, and Pleural Effusions

We measured K_f and wet-to-dry lung weights as indices of the lung permeability barrier in WT and TERT KO rats under normoxic and hyperoxic conditions (Fig. 14). K_f and wet-to-dry weights increased with hyperoxia in WT rats and TERT KO rats, consistent with previous reports. There were small differences that did not reach statistical significance in the K_f and wet-to-dry lung weights in TERT KO rats relative to those of the WT samples (P = 0.09 and 0.08, respectively). Subacute treatment with trehalose prevented the increase in wet-to-dry weight of TERT WT and KO rats exposed to hyperoxia. Pleural effusions were observed in hyperoxia-treated WT and KO rats.

Figure 14.

Microvascular permeability changes in wild-type (WT) and knockout (KO) rats. K_f (A), pleural effusions (B), and lung wet-to-dry weights (C) of lungs from WT and TERT KO rats maintained in normoxia or hyperoxia for 48 h. There were no differences in any of these endpoints between WT and KO rats in normoxia. Hyperoxia increased K_f, pleural effusions, and wet to dry weights of both WT and KO rats. The differences in K_f and wet to dry weights in WT and KO rats 48 h after hyperoxia trended to be increased but did not reach statistical significance. Subacute treatment with trehalose decreased wet to dry weights in hyperoxic TERT WT and KO rats. Pleural effusions were observed in WT and KO rats exposed to hyperoxia. Treatment with trehalose mitigated hyperoxia-induced collection of pleural effusion in WT and KO rats. A: *P < 0.02 vs. WT normoxia; n = 4–10 replicates. B: *P< 0.05 test vs. WT normoxia vehicle; n = 2–17 replicates. C: *P < 0.05 test vs. WT normoxia vehicle; n = 2–17 replicates.

DISCUSSION

We explored the role of autophagy/mitophagy and TERT in hyperoxic lung injury, given the clinical importance of extended exposure to high fractions of oxygen in patients with ARDS, including that caused by COVID-19. Lungs of rats exposed in vivo to hyperoxia exhibit activation of autophagy and mitophagy, as well as diminished mitochondrial function, as seen by changes in mitochondrial complex expression and MTT. Hyperoxia-induced mitochondrial dysfunction in cells is mitigated by activation of autophagy (trehalose or TSA) or stimulation of TERT expression (AGS 499). Similarly, MTT in lung tissue of rats exposed in vivo to hyperoxia can be improved by autophagy activation with trehalose. TERT expression is decreased in RLMVECs and lungs of WT and KO rats exposed to hyperoxia, and there are increased inflammatory changes in lungs of hyperoxic KO rats. TERT KO rats have altered mitochondrial proteostasis relative to that of WT cohorts based on greater hyperoxia-induced increases in TOMM20, no decrease in CII expression, no decrease in MTT, and no conversion of LC3B-I to II after hyperoxia exposure. The TERT KO rats also show dysfunction in normoxia based on lower MTT, which can be rescued by subacute treatment with trehalose to activate autophagy. Compromised MTT in hyperoxia-exposed RLMVECs can be rescued by stimulated expression of TERT. These changes in autophagy and mitochondrial activity have functional significance because hyperoxia-impaired TEER is mitigated by activation of autophagy or enhanced TERT expression. Activation of autophagy using trehalose mitigated hyperoxia-induced increases in wet-to-dry weight accumulation in WT and KO rats. These studies in a preclinical model of acute lung injury provide new mechanistic insights to guide better means to detect and/or treat complications of the life-saving therapy, hyperoxia.

Hyperoxic Effects on Mitochondrial Density

Cells challenged with ROS have been reported to exhibit more copies of _mtDNA. Liu et al. (25) observed that oxidative stress increased _mtDNA copy number in human leukocytes, with a simultaneous increase of three biomarkers of ROS exposure (the incidence of 4,977 bp deletion of _mtDNA, 8-OHdG content in leukocytes, and TBARS in plasma). Similarly, increased mitochondrial density is reported under other conditions of chronic stress, high glucose, and radiation. Kumari et al. (26) observed increased _mtDNA/_nucDNA ratios in the right ventricles of neonatal rat pups exposed to hyperoxia for 90 days, and increased _mtDNA mutations, as evidenced by decreased NADH dehydrogenase-4-to-NADH dehydrogenase-1 (ND4:ND1) ratio (26). Our data demonstrate increased TOMM20 protein in cells exposed to hyperoxia, as well as increases in lung tissue from KO rats relative to WT rats in normoxia. We also observed a greater hyperoxia-induced increase in TOMM20 in KO relative to WT rats. These data support hyperoxic injury to mitochondria and hence accumulation of inefficient organelles in RLMVECs. Defective mitochondrial importation and translocation of neuronal TOMM20 to the outer membrane has been reported to be associated with degraded CI activity in a murine model of neurogenerative disorder (27). Higher mitochondrial outer membrane protein densities in TERT KO lungs under normoxia and an insignificant increase with hyperoxia is consistent with greater oxidative stress at baseline in this strain.

Enhanced ROS and Mitochondrial Damage with Hyperoxia

Mitochondrial DNA damage, as detected by colocalization of the mitochondrial membrane protein TOMM20 and 8-oxo-dG, has been reported in cultured mouse lung epithelial cells exposed for 24 h to ${\text{FI}}_{{\text{O}}_2}$ 0.6 (28). Similarly, elevated levels of 8-oxo-dG were identified in the lungs of mice exposed to ${\text{FI}}_{{\text{O}}_2}$ 0.6 for only 2 h, indicating significant _mtDNA damage in the lungs of these animals. _mtDNA is known to be more sensitive to oxidative injury than _nucDNA in endothelial cells (29). There are _mtDNA hot spots including the Dloop known to be associated with oxygen radical damage (20). Based on these observations, we probed cultured RLMVECs and rat lungs exposed in vivo for _mtDNA damage. Figure 5 shows evidence of increases in the _mtDNA/_nucDNA ratio in hyperoxic cells but a decrease in this ratio in rat lungs. Though we do not see an increase in _mtDNA damage in lung tissue in TERT KO rats relative to that of WT cohorts in normoxia, we previously reported a significant increase in _mtDNA damage in hearts from KO rats compared with WT (9). The lack of differences between lesions/10 kb in lung tissue compared with heart tissue may be due to either contribution of cell types other than endothelia or compensatory mechanisms in vivo (4).

Autophagy Responses to Hyperoxia

Macroautophagy is a scalable process through which cells degrade dysfunctional components to maintain physiological homeostasis (30). As such, autophagy is generally considered to be a prosurvival pathway, the disruption of which, in the endothelium, results in dysfunction including loss of nitric oxide production (31). Hyperoxia is reported to promote proliferation of alveolar epithelial cells in neonatal mice (32), though the role of autophagy in this work was not established. Expression of Beclin-1, which initiates autophagy, is reported to be decreased in neonatal rat brains exposed to ${\text{FI}}_{{\text{O}}_2}$ 0.8 for 24 h (33), as we have observed in lungs. Tanaka et al. (34) described increased conversion of LC3B-I to LC3B-II in mouse lung epithelial cells exposed in vivo or in vitro to hyperoxia. Exposure of mice to ${\text{FI}}_{{\text{O}}_2}$ 0.3 (hyperoxia relative to normoxia) for 28 days did not change LC3B or p62 expression in cardiomyocytes, though it did result in increased apoptosis and _nucDNA fragmentation (35). Hyperoxia induced time-dependent conversion of LC3B-I to LC3B-II in homogenates of mouse lung exposed in vivo (48 and 72 h) to ${\text{FI}}_{{\text{O}}_2}$>0.95 and in cultured epithelial cells in vitro (34). The authors interpreted their data to imply that hyperoxia-induced LC3B activation confers cytoprotection in lung epithelial cells in a Fas apoptotic-dependent manner. Our data show that in response to hyperoxia, WT lungs exhibit decreased Beclin-1, LC3B-II, and ratio of LC3B-II/I relative to normoxic conditions, with no differences in LC3BII and LC3BI in TERT KO lungs. Particularly in the context of elevated p62, these data support altered autophagic flux or stalled autophagy in WT rat lungs. Autophagy turnover may also be greatly enhanced (e.g., LC3B-I to LC3B-II is accelerated but autophagosomes are degraded more quickly). However, without the use of a lysosomal inhibitor, we cannot definitively address this question. In light of the significantly elevated p62, we favor stalled autophagy. Understanding whether the defect in autophagy is due to stalled autophagy or enhanced autophagy turnover, as well as what defects are present in KO tissue that alter these processes, awaits further studies.

p62 is recognized as a reciprocal marker of autophagy and is widely used as a marker of autophagosome clearance (36). p62 can be recruited to the ubiquitinated mitochondria in CCCP-treated Parkin-positive cells (37–39). When autophagy is disrupted, p62 accumulates within ubiquitin-containing aggregates and in the cytoplasm. An initial rise in p62 seems beneficial for the cell, ensuring turnover of potentially harmful proteins by autophagy. But the sustained increases in p62, we reported in cultured endothelial cells (4) and now in lungs of WT and TERT KO rats exposed in vivo (Fig. 11G) provide support for a stalled autophagic process that results in overall increase of mitochondrial membrane protein/cell. We are not aware of other publications demonstrating impaired autophagy in hyperoxic pulmonary endothelial cells and intact rodent lungs linked to increased mitochondrial membrane protein.

The potential of mitochondria to regulate and be regulated by autophagy has recently received attention (40–42). Mammalian cells respond to an acute mitochondrial dysfunction by removing defective organelles with enhanced autophagosome formation and lysosomal hydrolytic capacity, thereby limiting ineffective ATP generation and increased ROS production. When mitochondrial defects persist, continuation of this response would result in catastrophic loss of energy-producing organelles. Accordingly, cells block autophagosomal flux and downregulate lysosomal function, leaving damaged (albeit semifunctional) mitochondria and impaired lysosomes. Thus, hyperoxic injury to mitochondria is potentially the primary stimulus for impaired autophagy and mitophagy we report here. We have previously shown that CI activity is decreased by ∼50% in mitochondria isolated from lungs of rats exposed for 48 h to hyperoxia (5, 43) and that hyperoxia promotes mitochondrial fission in cultured lung endothelial cells (4). The present study demonstrates that hyperoxic exposure results in no change in Parkin in WT or KO rat lung homogenates (Fig. 12), whereas Pink1 expression is increased in both WT and KO rats in response to hyperoxia. The full implications of these findings cannot be determined based on the data in this communication and await additional studies. However, these findings are consistent with increased mitochondrial proteins (TOMM20) and they may reflect adaptations of the mitophagy/autophagy systems to chronic mitochondrial stress in this strain. Impaired autophagy in hyperoxic lungs, which we describe in Fig. 11, is consistent with our observations of mitochondrial dysfunction and impaired MTT (Figs. 9 and 10).

Role of TERT as It Relates to Autophagy and ROS

A noncanonical role of TERT in preventing increased reactive oxygen species (ROS) and reduced ATP generation due to _mtDNA damage has been established (7, 44, 45). Insufficient autophagy or excessive levels of autophagy are both linked to cardiovascular and pulmonary disease (46, 47). Our group has demonstrated that coronary artery disease is associated with reduced TERT levels and activity (7, 9). We have recently shown that autophagy is responsible for the pathological switch in dilator mechanisms in arterioles from patients with coronary artery disease, acting downstream of telomerase as a common denominator of flow mediated dilation (8). Inhibition of TERT activity in murine fibroblasts reduces the conversion of LC3B-I to LC3B-II (48). Our data suggest TERT KO rats exhibit decreased conversion of LC3B-I to II relative to WT and fail to change the ratio of LC3B-II/I with hyperoxia (Fig. 7), consistent with stalled autophagy in these rats.

In atrial vessels from human subjects with coronary artery disease, the increase in _mtROS, altered endothelial function, and diminished nitric oxide can be mitigated by increased TERT, which upregulates mitochondrial antioxidant enzymes (49), reduces _mtROS production, and reverses the impaired phenotype observed in coronary vessels from CAD human subjects (7). Our new data demonstrate lower MTT activities in TERT KO rats relative to WT in normoxia (baseline, unstressed condition), which is consistent with a chronic mitochondrial stress attributable to deficient functional TERT. Hyperoxia decreased TERT expression in RLMVECs and lungs of WT hyperoxic rats exposed in vivo to hyperoxia (Fig. 4). Pharmacological stimulation of TERT expression diminished hyperoxic injury to mitochondria in cultured RLMVECs based on MTT studies (Fig. 9). Upregulation of TERT expression using AGS 499 cannot overcome the inhibition of autophagy with BAFA1. Similarly, stimulation of autophagy with trehalose does not overcome inhibition of TERT activity with BIBR 1532. We cannot exclude nuclear actions of TERT in protection from hyperoxia-induced autophagy; DNA repair by oxoguanine glycosylase 1 is reported to promote autophagy in lungs of mice exposed to hyperoxia (50) via cytokine release. Taken together, our observations suggest that TERT activity and active autophagy are needed for optimal mitochondrial function in RLMVECs. Please see Fig. 15 for our hypothesized signaling pathways that activate TERT and autophagy after hyperoxia.

Figure 15.

This schematic depicts our observations regarding hyperoxic lung injury (central column), the cellular processes affected (left column), and the effect of TERT on these processes (right column). Hyperoxia initiates autophagy, as evidenced by decreasing the Beclin-1 and LC3B-II/I ratios and increasing p62. Increased PINK1, increased activating substances for TLR9, decreased MTT, and increased TOMM20 expression support hyperoxia-associated mitochondrial injury in exposed endothelial cells and lung tissue. These changes are associated with neutrophil influx and increased myeloperoxidase (MPO), evidence of inflammation. Functional consequences of hyperoxia include diminished monolayer integrity and increases in wet-to-dry lung weight and pleural effusions. Activation of autophagy protects from hyperoxia-induced decrease in MTT and provides protection against loss of barrier function with hyperoxia. TERT contributes to maintenance of mitochondrial and barrier function in hyperoxia. MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide.

Functional Significance

Hyperoxic lung injury is characterized by increased microvascular permeability in vivo, which can be quantified by measurements of K_f (filtration coefficient), wet-to-dry weights of lungs, and pleural effusions (13, 51). TEER affords a quantitative measure of endothelial function and is an in vitro model of hyperoxic injury that is sensitive to hyperoxia (4). In the present study, we tested the potential of TERT and autophagy stimulators to protect against hyperoxic-induced reduction in endothelial monolayer resistance. We observed that activation of either system preserved TEER, consistent with the functional significance of these signaling systems and suggesting converging activities. To test the significance of TERT KO in hyperoxic exposure in vivo, we measured K_f and wet-to-dry weights of lungs of WT and KO rats exposed to hyperoxia for 48 h. Both K_f and wet-to-dry weights show a trend to increased injury in KO rats relative to WTs but did not reach statistical significance. Subacute treatment with trehalose decreased wet-to-dry weights of hyperoxic WT and TERT KO rats, providing a possible link between autophagy and TERT in hyperoxic injury. Of translational interest, treatment of human coronary arterioles with trehalose for only 15–20 h restores nitric oxide-mediated dilation, suggesting that short-term treatments modify functional endpoints (8) Although the cell culture data did not perfectly match in vivo results, there may be several reasons for this, including lower sensitivity in TERT KO rats induced by compensatory adaptations in vivo. Our TERT KO rats have globally nonfunctional TERT, and TERT is likely one of many signaling systems that modify microvascular permeability in the setting of hyperoxia. TERT activity in cell types other than pulmonary endothelial cells, or elsewhere besides the mitochondria, may have different effects than those we observed in isolated RLMVECs. We know that histologically there were no differences in the lungs of WT and KO rats under normoxic conditions but increased inflammatory changes in hyperoxia, which are augmented in KO rats.

In conclusion, we have shown that hyperoxia impairs mitochondrial function, autophagy, and mitophagy in pulmonary endothelial cells in culture and in vivo in a manner that changes mitochondrial membrane protein density, suggests cross talk between TERT and autophagy signaling, and implies the requirement for both to withstand this stress optimally.

SUPPLEMENTAL DATA

Supplemental material: https://doi.org/10.6084/m9.figshare.16540323.v1.

GRANTS

This work was supported by National Institutes of Health (NIH) Grants HL116530 (to E. R. Jacobs), HL133029 (to A. M. Beyer), HL1202209 (to S. H. Audi), HL129209, and T32GM089586; American Heart Association Grant 20POST35050017 (to W. E. Hughes); Veterans Affairs Merit Review Award 1I01BX001681 (to E. R. Jacobs); NIH Grants AI101898, AI107305, and OD018306 (to M. Medhora); and the Department of Radiation Oncology, Medical College of Wisconsin.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

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