OTULIN protects hyperoxia-induced neonatal lung injury and modulates mitochondrial protein OPA1 in association with the E3 ubiquitin ligase RNF31

Abstract

The deubiquitinating enzyme OTULIN has been implicated in the development of lung injury, and regulating its expression may either exacerbate or alleviate pulmonary inflammatory damage. In this study, we aimed to investigate the role of deubiquitinating enzyme OTULIN in hyperoxia-induced lung injury and the underlying mechanisms involved. A bronchopulmonary dysplasia (BPD) model was established by exposing neonatal mice to a hyperoxic environment, and the effects of regulating OTULIN expression on mitochondrial homeostasis in pulmonary epithelial cells were further examined under hyperoxic conditions. In addition, we investigated the mechanisms through which OTULIN regulates mitochondrial-associated proteins and the ubiquitination mechanisms of differential mitochondrial protein OPA1. The results showed that hyperoxia induced significant lung injury in neonatal mice and was accompanied by upregulation of OTULIN expression. Additionally, hyperoxia disrupted mitochondrial homeostasis in neonatal mice lung tissue, as observed by a reduction in mitochondrial number and increased mitochondrial fusion and autophagy. After hyperoxia exposure, overexpression of OTULIN significantly reduced mitochondrial reactive oxygen species (ROS) levels in alveolar epithelial cells, maintained mitochondrial membrane potential, and promoted mitochondrial homeostasis. Mechanistically, OTULIN was found to directly interact with OPA1 and regulate its ubiquitination status. The E3 ubiquitin ligase RNF31 was identified as a key regulator of OPA1 stability, with knockdown of RNF31 reducing OPA1 levels. Moreover, OTULIN regulated the expression of both OPA1 and RNF31 and affected the stability of OPA1 and mitochondrial function through RNF31-dependent mechanisms. In vivo experiments further showed that knockdown of OTULIN aggravated hyperoxia-induced lung injury in neonatal mice, characterized by alveolar simplification, increased fibrosis, and further impairment of mitochondrial function, whereas overexpression of OTULIN alleviated these pathological changes. In conclusion, deubiquitinating enzyme OTULIN protected hyperoxia-induced neonatal lung injury and modulates mitochondrial protein OPA1 in association with the E3 ubiquitin ligase RNF31. These findings provide new insights into the pathogenesis of BPD and highlight the therapeutic potential of targeting the OTULIN/RNF31–OPA1 axis.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s11658-026-00946-4.

Introduction

Bronchopulmonary dysplasia (BPD) is a debilitating disease affecting premature infants and is caused by lung injury that disrupts the normal development of alveoli and pulmonary vasculature [1]. Hyperoxia-induced inflammation and tissue damage are key steps leading to BPD in immature lungs [2]. Advances in neonatal intensive care have improved the survival rate of premature infants but have not reduced the incidence of BPD, thereby imposing a substantial medical and economic burden on families and society [3]. However, the etiology and pathogenesis of BPD remain incompletely understood [4]. The pathogenesis of BPD is multifactorial, and the clinical phenotype is heterogeneous [5]. Therefore, further exploration of the mechanisms underlying BPD and the identification of new effective treatment strategies are of great significance.

Mitochondria are highly dynamic cellular components, and their functional flexibility depends on continuous processes of membrane fusion and fission, which are crucial for maintaining mitochondrial integrity, turnover, and overall function [6]. Maintaining an appropriate equilibrium between fusion and fission, collectively known as mitochondrial dynamics, is essential for preserving mitochondrial homeostasis [7]. Mitophagy is a mechanism for the selective elimination of damaged mitochondria, thereby reducing the production of mitochondrial reactive oxygen species (mtROS) and promoting cell survival [8]. Studies have shown that mitochondrial dysfunction occurs in neonatal lung injury [9]. In hyperoxia-induced injury of type II alveolar epithelial cells, mitochondrial homeostasis is affected by Parkin-regulated mitophagy [10]. Furthermore, studies using various lung injury models have indicated that regulating mitochondrial components or functions could improve lung injury [11–13]. However, the regulatory pathways related to mitochondrial quality control, including mitochondrial dynamics and mitophagy, in hyperoxia-induced lung injury are not yet fully understood.

The process of ubiquitination, an important posttranslational protein modification, is essential for regulating protein stability [14]. E3 ubiquitin ligases are instrumental in the precise identification and regulation of substrates for ubiquitination reactions [15]. In contrast, deubiquitinating enzymes remove ubiquitin from target proteins and regulate protein stability, whereas E3 ubiquitin ligases mediate the conjugation of ubiquitin to target proteins, usually leading to protein degradation [16]. Studies have shown that mitochondrial homeostasis is regulated by ubiquitination, and deubiquitinating enzymes such as USP14, OTUD1, PSMD14, and OTULIN could affect mitochondrial damage by regulating the deubiquitination of related proteins [17–21]. Additionally, ubiquitination-related proteins, including deubiquitinating enzymes such as USP14 and OTULIN, are also involved in lung injury development, and regulating their expression could exacerbate or alleviate pulmonary inflammatory damage [22, 23]. Therefore, in this study, we investigated the mechanisms by which the deubiquitinating enzyme OTULIN and E3 ubiquitin ligases regulate mitochondrial homeostasis in BPD.

Against this background, this study conducted in vivo and in vitro experiments to elucidate the mechanism of the deubiquitinating enzyme OTULIN and E3 ubiquitin ligase in hyperoxia-induced lung injury., providing insights into the pathogenesis of neonatal BPD, which could facilitate the development of promising therapeutic strategies.

Materials and methods

Animals

Specific-pathogen-free (SPF)-grade newborn C57BL/6 pups (both sexes, within 12 h of birth) were purchased from Hunan SJA Laboratory Animal Co., Ltd. After arrival, litters delivered on the same day were pooled and randomly assigned to either the Air or Hyperoxia group without sex preselection. Pups in the Hyperoxia group were housed in an enclosed Plexiglass chamber (temperature maintained at 22–27 °C, humidity between 50% and 70%). They were subjected to an environment with 95% O₂ from birth until day 14. The hyperoxic condition was maintained with a continuous oxygen supply at a rate of 4 L/min, and O₂ levels were continuously monitored using an oxygen sensor. Normoxia and hyperoxia conditions were alternated every 24 h to limit oxygen toxicity. For time-course analysis of hyperoxia exposure, pups were subjected to two distinct exposure durations: from birth until postnatal day 7 (P7) or day 14 (P14). The Air group was kept in normal indoor air (21% O₂).

Mice in the Hyperoxia + short hairpin RNA (shRNA) negative control (sh-NC) and Hyperoxia + sh-OTULIN groups were subjected to hyperoxia exposure and received intraperitoneal injections of 30 μL sh-NC /sh-OTULIN lentivirus (1 × 10⁸ TU/mL) every 3 days [24]. Additionally, mice in the Hyperoxia + overexpression (oe)-NC and Hyperoxia + oe-OTULIN groups were subjected to hyperoxia exposure and received intraperitoneal injections of 30 μL oe-NC/oe-OTULIN lentivirus (1 × 10⁸ TU/mL) every 3 days [24]. Pups received the first intraperitoneal injection of lentivirus or vehicle at P3. Lentivirus (sh-OTULIN or sh-NC, oe-OTULIN or oe-NC; 1 × 10⁸ TU/mL, 30 μL per pup) was administered i.p. every 3 days (P3, P6, P9, and P12) for a total of four injections. Mice in the Air and Hyperoxia groups were injected with 0.9% NaCl solution.

Lung tissues were collected on P14 (within 2 h after the final virus injection). The pups were anesthetized via intraperitoneal injection of sodium pentobarbital (150 mg/kg). Immediately thereafter, the abdominal cavity was opened, and the right lung was excised and placed into an RNase-free cryotube. After rapid freezing in liquid nitrogen, the tissue was stored at −80 °C for subsequent quantitative real-time polymerase chain reaction (qRT-PCR) and Western blot analyses [25]. Next, 4% paraformaldehyde was slowly injected via the left bronchus until the lung apex was inflated. The lung was then placed into an embedding cassette and immersed in 4% paraformaldehyde for overnight fixation (at least 24 h to ensure complete fixation [26]), followed by subsequent experimental analysis.

Extraction of primary mouse lung epithelial cells

The C57BL/6 mouse was fully sterilized with 75% ethanol. After fixation in the supine position, the skin and thoracic cavity were incised to fully expose the heart and lungs. The right auricle was cut open, and a blunt-ended needle was inserted into the right ventricle or main pulmonary artery. Approximately 5–10 mL of precooled sterile Dulbecco’s phosphate-buffered saline (DPBS, calcium- and magnesium-free) was slowly perfused until the lung color changed from pink to uniformly white. The intact lung tissue was removed and placed in a Petri dish containing a small amount of Dulbecco’s modified Eagle medium (DMEM)/F12 medium. Using fine forceps, the trachea, main bronchi, and visible blood vessels were carefully dissected and removed. The remaining lung parenchymal tissue was transferred to a Petri dish containing a small volume of medium and minced with two crossed sterile surgical blades until a paste-like consistency was achieved. Collagenase II was then added, and the mixture was digested on a shaker at 37 ℃ for 30 min. DMEM/F12 complete medium was added to terminate the digestion, and the mixture was gently pipetted 50–100 times using a large-bore pipette (e.g., a 10-mL pipette). The digested tissue was sequentially filtered through 70-μm and 40-μm cell strainers to remove undigested tissue clumps and debris. The filtrate was collected and centrifuged at 300–400 × g for 10 min at 4 ℃. The supernatant was discarded, and the cells were resuspended in complete medium.

The cell suspension was seeded into uncoated tissue culture dishes and incubated at 37 °C in a 5% CO₂ incubator for 2 h. The nonadherent cell suspension was carefully collected, centrifuged, and resuspended in sorting buffer (e.g., DPBS containing 2 mM ethylenediaminetetraacetic acid (EDTA) and 0.5% bovine serum albumin (BSA)) for cell counting. According to the manufacturer’s instructions for the magnetic bead sorting kit, the cells were sequentially incubated with anti-CD45 microbeads and anti-CD16/32 microbeads at 4 °C. The cell suspension was then passed through an LD column, and the eluted negative cells were collected by centrifugation. After discarding the supernatant, the cells were resuspended in specialized complete medium and seeded into T25 culture flasks, which were then incubated at 37 °C in a 5% CO₂ incubator for culture [27].

Cell culture and treatment

Human alveolar epithelial A549 cells (AW-CCH011, Abiowell) were cultured in F-12 K medium (AW-M004, Abiowell) supplemented with 10% fetal bovine serum and 1% penicillin–streptomycin. Mouse alveolar epithelial MLE-12 cells (ZQ0470, ZQXZBIO) were cultured in MLE-12 -specific medium (CM-0680, Procell). To construct a hyperoxia cell model, cells in the Control and hyperoxia (Hyp) groups were exposed to air (21% O₂) or hyperoxia (85% O₂) for 48 h [28]. To knock down or overexpress OTULIN, A549 and MLE-12 cells were transfected with sh-OTULIN or overexpression (oe)-OTULIN plasmids, respectively. After 48 h of transfection, the cells were subjected to hyperoxia exposure. The groups were designated as follows: Control, Hyp, Hyp + sh-NC, Hyp + sh-OTULIN, Hyp + oe-NC, and Hyp + oe-OTULIN. For RNF31 knockdown, the groups were designated as sh-NC and sh-RNF31. For OPA1 mutation experiments, the groups were designated as OPA1–wt (full length: 1–960), OPA1-mut#1 (1–469/△ 470–960), and OPA1-mut#2 (1–290/△ 291–960). Furthermore, for combined OTULIN and RNF31 knockdown, A549 and MLE-12 cells were transfected with sh-OTULIN and sh-RNF31 for 48 h before hyperoxia exposure. The groups were designated as follows: Control, Hyp, Hyp + sh-NC, Hyp + sh-OTULIN, Hyp + sh-OTULIN + sh-NC, and Hyp + sh-OTULIN + sh-RNF31. Additionally, OTULIN knockdown experiments were conducted in primary mouse alveolar epithelial cells, with the following groups: Control + sh-NC, Control + sh-OTULIN#1, Control + sh-OTULIN#2, hyperoxia + sh-NC, hyperoxia + sh-OTULIN#1, and hyperoxia + sh-OTULIN#2.

The interference sequences were as follows: Mouse sh-OTULIN#1, a short hairpin RNA targeting OTULIN, sequence 5′-GTGGAAACTTGGACTGAAATT-3′, catalog no. HG-SM001416048 (HonorGene); Mouse sh-OTULIN#2, sequence 5′-GACGGTAAACTGGTTTTACTT-3′, catalog no. HG-SM001416048 (HonorGene); Mouse sh-RNF31, a short hairpin RNA targeting RNF31, sequence 5′-ACGGTTGTATGGCTATACTGA-3′, catalog no. HG-SM057595 (HonorGene); Mouse sh-NC, a nontargeting negative control shRNA, sequence 5′-TTAACCGTAGACGGAGAAGAC-3′; Human sh-OTULIN, sequence 5′-GGGCTATGAAGAGGTTTCTCA-3′, catalog no. HG-SH138348 (HonorGene); Human sh-RNF31, sequence 5′-GCCTTTTACGCCAAGAATAAA-3′, catalog no. HG-SH017999 (HonorGene); and Human sh-NC, sequence 5′-TGCCCAAACCCCAACCCAACC-3′. The overexpression catalog numbers were as follows: Mouse oe-OTULIN: catalog no. HG-MO001416048 (HonorGene); Human oe-OTULIN: catalog no. HG-HO138348 (HonorGene). OPA1–wt was a full-length (1–960 aa) expression plasmid of OPA1, catalog no. HG-HO015560 (HonorGene); OPA1-mut#1 was a truncated variant (1–469 aa, Δ470–960) of OPA1, catalog no. HG-HO015560–1 (HonorGene); and OPA1-mut#2 was a truncated variant (1–290 aa, Δ291–960) of OPA1, catalog no. HG-HO015560–2 (HonorGene).

Hematoxylin–eosin (HE) staining

HE staining was employed to examine the structural characteristics of mouse lung tissue. The tissue sections were dried at 60 °C for 12 h. They were first immersed in xylene for 20 min, and this step was repeated three times. The sections were then rehydrated through a graded ethanol series of 100%, 100%, 95%, 85%, and 75%, with each step lasting 5 min, followed by washing in distilled water for 5 min. The sections were stained with hematoxylin for 1–10 min, rinsed in distilled water, and treated with phosphate-buffered saline (PBS) to restore the blue color. Eosin staining was then performed for 1–5 min, followed by another rinse in distilled water. Subsequently, the sections were dehydrated through an ascending ethanol series (95% to 100%) for 5 min at each step or air-dried. Finally, the sections were cleared in xylene for 10 min, and this step was repeated twice, before mounting with neutral resin and microscopic examination.

Lung injury assessment

On the basis of a previous study [29], lung tissue architecture was evaluated by determining the mean linear intercept (MLI) and radial alveolar count (RAC) in HE-stained sections. From the center of each respiratory bronchiole, four perpendicular radial lines were drawn extending to the adjacent acinar margin. The number of alveolar septa transected by each line was counted, and the average value across all lines and bronchioles was calculated as the RAC for each sample [30]. The MLI was determined using a standard point- and intersection-counting method [26, 31]. A coherent test system of parallel lines with known length was randomly superimposed on 20–30 non-overlapping, randomly selected microscopic fields per lung. The total length of the test lines and the total number of intersections between test lines and alveolar septa were recorded. MLI was calculated as: MLI = total length of test lines/total number of intersections. To ensure accuracy, large airways and vessels were excluded from analysis, and all measurements were performed in a blinded manner.

Masson staining

Masson staining was applied to identify fibrotic areas within mouse lung samples. The sections were heated at 60 °C for more than 12 h, followed by deparaffinization and rehydration. A nuclear counterstain (Weigert’s iron hematoxylin) was evenly applied to the tissue sections and incubated for 10 s to 1 min to visualize cell nuclei. The staining solution was thoroughly washed off with tap water, followed by a brief rinse in distilled water. Subsequently, the sections were immersed in a mildly alkaline solution, such as PBS (pH 7.2–7.6) or ammonia water, for 5–10 min to restore the blue color of the nuclei. An appropriate amount of connective tissue stain was then applied to fully cover the entire section, and incubated for 2–5 min before being completely washed away. The sections were briefly treated with a differentiating solution for approximately 30 s, after which the solution was removed without rinsing. A counterstaining solution was then evenly applied and incubated for 3–8 min, followed by thorough rinsing with anhydrous ethanol. Finally, the sections were air-dried, cleared with xylene, mounted with neutral mountant, and examined under a microscope.

qRT-PCR

qRT-PCR was employed to assess OTULIN expression in lung epithelial cells, as well as OTULIN, BNIP3, BNIP3L, FUNDC1, OPA1, PINK, PARK2, and TOM20 expression in primary mouse lung epithelial cells. Total RNA was initially isolated using a TRIzol total RNA extraction kit (15,596,026, Thermo), after which RNA concentration and purity were determined. Subsequently, messenger RNA (mRNA) was reverse-transcribed into complementary DNA (cDNA) using an mRNA reverse transcription kit (CW2569, CWBIO). Relative gene expression was quantified using Ultra SYBR Mixture (CW2601, CWBIO) on an ABI 7900 platform. β-actin was used as the endogenous control, and expression levels were calculated using the 2^−ΔΔCt method. The primer sequences are listed in Table 1.

Table 1.

The primers used in this study

Names	Sequences (5′–3′)	Accession no.
H-OTULIN-F	TGAGTCGGGGGACTATGCC	NM_138348
H-OTULIN-R	AATCTCGGTTCTGATGCCCC
H-β-actin-F	ACCCTGAAGTACCCCATCGAG	NM_001101
H-β-actin-R	AGCACAGCCTGGATAGCAAC
M-OTULIN-F	CGAGGAGGACATGTACCGTG	NM_001013792
M-OTULIN-R	AGTTATCACCACGGACTCGC
M-BNIP3-F	AAATTAAAGGGTGCGTGCGG	NM_009760
M-BNIP3-R	CAAAGTGGGGTTCGTGGGTA
M-FUNDC1-F	TGGCTGGTGTGCAGGATTTT	NM_001313745
M-FUNDC1-R	AGTCTGTTGCCTGAAGAAGAAGG
M-TOM20-F	CAAACACCGGGATGCAAACA	NM_024214
M-TOM20-R	GATGGAACACCCCAGAGACG
M-BNIP3L-F	AAGGGAAGGAACAACTTCACA	NM_009761
M-BNIP3L-R	ATGGAAGACGAGGAAGGAACG
M-PARK2-F	ACACATAGTACAGAGACCACGGA	NM_016694
M-PARK2-R	CAGCCCCACAGAGTCCACC
M-OPA1-F	TTCTTCACTGCAGGTCCCAAAT	NM_001199177
M-OPA1-R	TCTGACACCTTCCTGTAATGCTTG
M-PINK1-F	CTCAAGTCCGACAACATCCTT	XM_036164400
M-PINK1-R	CCATTGCCACCACGCTCT
M-β-actin-F	ACATCCGTAAAGACCTCTATGCC	NM_007393
M-β-actin-R	TACTCCTGCTTGCTGATCCAC

Western blot

Western blot was performed to evaluate the expression levels of OTUD1, PSMD14, USP14, OTULIN, TOM20, the mitochondrial fusion-associated protein OPA1, mitochondrial fission-related proteins (DRP1 and phosphorylated DRP1 at Ser616), proteins associated with mitochondrial autophagy (PINK1, PARK2, BNIP3L, BNIP3, FUNDC1, and LC3), the E3 ubiquitin ligase RNF31, and RBCK1 in lung tissues or cells. Samples from different groups were lysed with radioimmunoprecipitation assay (RIPA) buffer (AWB0136, Abiowell) to extract total protein. The proteins were separated by sodium dodecyl sulfate (SDS) polyacrylamide gel electrophoresis (PAGE) and transferred onto a nitrocellulose membrane. The membrane was then blocked with 5% milk at room temperature for 1.5 h, followed by incubation with primary antibodies at 4 °C overnight. The primary antibodies included OTUD1 (29,921-1-AP, 1:2000, Proteintech), PSMD14 (12,059-1-AP, 1:1000, Proteintech), USP14 (14,517-1-AP, 1:5000, Proteintech), OTULIN (27,392-1-AP, 1:2000, Proteintech), TOM20 (11,802-1-AP, 1:2000, Proteintech), OPA1 (ab157457, 1:3000, Abcam), DRP1 (ab184247, 1:1000, Abcam), p-DRP1 Ser616 (ab314755, 1:1000, Abcam), PINK1 (AWA54820, 1:1000, Abiowell), PARK2 (ab77924, 1:2000, Abcam), BNIP3L (ab109414, 1:5000, Abcam), BNIP3 (ab109362, 1:1000, Abcam), FUNDC1 (AWA51295, 1:2000, Abiowell), LC3 (14,600-1-AP, 1:3000, Proteintech), RNF31 (16,289-1-AP, 1:1000, Proteintech), RBCK1 (ab309104, 1:1000, Abcam), and β-actin (66,009-1-Ig, 1:5000, Proteintech). The membrane was then incubated with horseradish peroxidase (HRP)-conjugated goat anti-mouse immunoglobulin G (IgG, SA00001-1, 1:5000, Proteintech) or HRP-conjugated goat anti-rabbit IgG (SA00001-2, 1:6000, Proteintech) at room temperature for 1.5 h. Following this, the membrane was incubated with enhanced chemiluminescence (ECL) reagent (AWB0005, Abiowell) for 1 min and then subjected to imaging for detection. β-Actin served as the loading control, and relative protein expression levels were quantified using Quantity One 4.6.6 software. The full uncropped blot images are provided in Supplementary File S1.

Immunofluorescence (IF)

IF staining was performed to detect TOM20 expression in lung tissue samples. The sections were heated at 60 °C for 12 h and then deparaffinized through xylene -to -water transitions, after which antigen retrieval was performed. The sections were incubated in sodium borohydride solution at room temperature for 30 min, followed by rinsing in water for 5 min. Subsequently, they were briefly immersed in 75% ethanol at room temperature for 5 min. The sections were then treated with Sudan Black staining solution at room temperature for 5 min and rinsed in water for 3 min. A blocking solution containing 10% normal serum and 5% BSA was applied for 1 h. The primary antibody against TOM20 (11,802-1-AP, 1:50, Proteintech) was then applied and incubated at 4 °C overnight. The sections were rinsed with PBS for 5 min, and this step was repeated three times. The secondary antibody, Goat anti-Rabbit IgG (H + L) Secondary Antibody (AWS0005a, Abiowell), was applied and incubated at 37 °C for 90 min, followed by another series of rinses with PBS for 5 min, repeated three times. The sections were then counterstained with 4′,6-diamidino-2-phenylindole (DAPI) working solution at 37 °C for 10 min, rinsed with PBS for 5 min, and this rinse step was also repeated three times. Finally, the sections were mounted with buffered glycerol and examined under a fluorescence microscope.

IF staining was performed to detect KRT8 expression in primary mouse lung epithelial cells and the levels of TOM20 and OPA1 in A549 and MLE-12 cells. The cells on the slides were fixed with 4% paraformaldehyde for 30 min, after which they were washed with PBS for 10 min, and this process was repeated six times. The slides were then treated with an immunostaining permeabilization solution containing Triton X-100 (P0096-100 ml, Beyotime) at 37 °C for 5 min. A 5% BSA blocking solution was applied to the slides and incubated at 37 °C for 60 min. The slides were incubated overnight at 4 °C with primary antibodies against KRT8 (17,514-1-AP, 1:50, Proteintech), TOM20 (11,802-1-AP, 1:50, Proteintech), and OPA1 (27,733-1-AP, 1:50, Proteintech). A CoraLite488-conjugated Goat Anti-Rabbit IgG (H + L) secondary antibody (SA00013-2, Proteintech) was added and incubated at 37 °C for 60 min. The slides were stained with DAPI working solution at 37 °C for 10 min. Finally, the slides were mounted with Fluoromount-G^® (0100-01, SouthernBiotech) and protected from light until examination under a fluorescence microscope.

IF staining was conducted to analyze the co-localization of OTULIN with differential mitochondrial-associated proteins (OPA1, PINK1, PARK2, and BNIP3L) in A549 and MLE-12 cells using a double-label multiplex immunofluorescence kit (AWI0702, Abiowell). The slides were fixed with 4% paraformaldehyde for 30 min and then washed with PBS for 10 min, and this procedure was repeated six times. The slides were then permeabilized with an immunostaining solution containing Triton X-100 at 37 °C for 5 min. Subsequently, 3% hydrogen peroxide solution was applied, and the slides were incubated in the dark at room temperature for 25 min, followed by washing three times with PBS for 5 min each. A 5% BSA blocking solution was added and incubated at 37 °C for 60 min. The slides were then incubated overnight at 4 °C with the appropriately diluted primary antibody OTULIN (ab151117, 1:100, Abcam). Subsequently, 50 µL of HRP-conjugated secondary antibody (SA00013-2, Proteintech) was applied and incubated at 37 °C for 60 min. The slides were reacted with the first fluorescent dye, TSA570, for 10 min, followed by an antibody elution step. The slides were again treated with 3% hydrogen peroxide solution, incubated in the dark at room temperature for 25 min, and washed three times with PBS for 5 min each. A 5% BSA blocking solution was reapplied and incubated at 37 °C for 60 min, followed by overnight incubation at 4 °C with diluted primary antibodies against OPA1 (27,733-1-AP, 1:100, Proteintech), PINK1 (ab216144, 1:200, Abcam), PARK2 (14,060-1-AP, 1:150, Proteintech), and BNIP3L (ab8399, 1:100, Abcam). Then, 50 µL of HRP-conjugated secondary antibody (SA00013-2, Proteintech) was added and incubated at 37 °C for 30 min, followed by reaction with another fluorescent dye, 520, for 10 min. The slides were stained with DAPI working solution at 37 °C for 10 min. Finally, the slides were mounted with Fluoromount-G^® (0100-01, SouthernBiotech) and kept in the dark or examined under a fluorescence microscope.

Immunohistochemistry (IHC)

IHC was utilized to assess OTULIN expression in lung tissue. Sections were heated at 60 °C for 12 h, deparaffinized, and subjected to antigen retrieval. A 1% hydrogen peroxide solution was added and incubated at room temperature for 15 min to inactivate endogenous enzymes. The sections were washed with PBS for 3 min, and this step was repeated three times. Appropriately diluted primary antibody against OTULIN (ab151117, 1:200, Abcam) was applied and incubated overnight at 4 °C. The sections were then rinsed with PBS for 5 min, and this step was repeated three times. A secondary antibody (AWI0629, Abiowell) was added and incubated at 37 °C for 30 min, followed by another round of rinsing with PBS for 5 min, repeated three times. A prepared 3,3′-diaminobenzidine (DAB)− chromogen working solution (50–100 μL) was applied and incubated at room temperature for 1–5 min, with the reaction monitored under a microscope, followed by rinsing with distilled water. The sections were counterstained with hematoxylin for 5–10 min, rinsed with distilled water, and treated with PBS to restore the blue color. The sections were dehydrated through a graded series of alcohol solutions (60–100%) for 5 min each. Thereafter, the sections were placed in xylene for 10 min twice and then mounted with neutral resin for microscopic observation. Integrated optical density (IOD) analysis was performed using Image-Pro Plus 6.0 software, and OTULIN levels were expressed as the average optical density (IOD of the positive area in the field of view/tissue area in the field of view).

Mitochondrial reactive oxygen species (ROS)

Flow cytometry was applied to analyze mitochondrial ROS levels in lung tissue. Lung tissue was minced into approximately 0.5 mm³ using surgical scissors and transferred to a centrifuge tube containing trypsin digestion solution. The tissue was incubated in a 37 °C cell culture incubator for 1 h with intermittent mixing. The cells were passed through a filter and transferred into a fresh centrifuge tube, followed by centrifugation at 1500 rpm for 5 min to pellet the cells. An appropriate volume of red blood cell lysis buffer was added to lyse any remaining red blood cells. After centrifugation at 1500 rpm for another 5 min, the cell pellet was collected and washed twice with PBS. 2′,7′-Dichlorodihydrofluorescein diacetate (DCFH-DA, stock concentration 10 mM) was diluted 1:1000 in serum-free culture medium to obtain a final concentration of 10 µM. For treated cells, the culture medium was removed and replaced with an appropriate volume of diluted DCFH-DA solution to ensure full coverage of the cells. The cells were then incubated at 37 °C in a cell culture incubator for 20 min. After incubation, the cells were washed three times with serum-free culture medium to thoroughly remove extracellular DCFH-DA. The cells were subsequently collected by trypsinization, and mitochondrial ROS levels were analyzed using a flow cytometer (A00-1-1102, Beckman).

Mitochondrial ROS were also analyzed by MitoSOX fluorescence staining. The stock solution was diluted with preheated serum-free cell medium to prepare a 1 μM MitoSOX Red working solution. The growth medium was aspirated from the cells, and an appropriate volume of the diluted reagent solution was added to ensure that the cells were completely covered. The cells were then incubated at 37 °C for 5–10 min. Following incubation, the cells were washed three times with a serum-free medium to remove residual reagent. Trypsin was used to digest the cells, and the cells were collected. Mitochondrial ROS levels were then detected using a fluorescence microscope.

Mitochondrial membrane potential

The mitochondrial membrane potential in lung tissue was assessed by flow cytometry using a procedure similar to that used for the detection of mitochondrial ROS in lung tissue. After centrifugation at 1500 rpm for 5 min, the cells were collected and washed twice with PBS. Subsequently, an appropriate volume of 10 µM JC-1 staining solution was added, and mitochondrial membrane potential was measured by flow cytometry.

Mitochondrial membrane potential was also determined using JC-1, a ratiometric cationic carbocyanine dye that accumulates in mitochondria in a mitochondrial membrane potential-dependent manner. When mitochondrial membrane potential is high, JC-1 forms red-fluorescent J-aggregates (excitation 535 nm/emission 590 nm), whereas when mitochondrial membrane potential collapses, the dye remains as green-fluorescent monomers (excitation 485 nm/emission 530 nm). Therefore, a decrease in the red/green fluorescence ratio indicates mitochondrial depolarization and serves as an early marker of mitochondrial damage. Briefly, cells were harvested and resuspended in 0.5 mL culture medium. An equal volume (0.5 mL) of JC-1 staining solution (200 µM, C2006, Beyotime) was added, gently mixed, and incubated at 37 °C for 20 min. After two washes with JC-1 staining buffer, the cells were resuspended in 2 mL fresh medium. Images were captured immediately under a fluorescence microscope (IX83, Olympus) using TRITC (red) and FITC (green) filters. The red/green fluorescence ratio was quantified using ImageJ and are presented in arbitrary units.

Bioinformatics analysis

The GSE25286 dataset was downloaded from the Gene Expression Omnibus (GEO) database and converted into an expression matrix, and the corresponding group information was obtained. The expression matrix was processed, and differential expression was analyzed using the R language. OTULIN-related genes were screened by Pearson correlation analysis (P < 0.05, |cor|> 0.5), and a total of 2982 genes were obtained. According to the screening results from the previous step, the clusterProfiler R package was used to perform Gene Ontology (GO) enrichment analysis.

Co-immunoprecipitation (Co-IP)

On the basis of previous research [32], total cellular protein was extracted, and Co-IP was used to detect whether OTULIN interacted with OPA1, PINK1, PARK2, and BNIP3L in cells. Using the OTULIN antibody as the bait protein, Western blot was performed after IP to detect the expression of OTULIN, OPA1, PINK1, PARK2, and BNIP3L. In addition, total cellular protein was extracted, and Co-IP was used to detect whether OPA1 interacted with RNF31 and RBCK1 in cells. Using the OPA1 antibody as the bait protein, Western blot was performed after IP to detect the expression of OPA1, RNF31, and RBCK1. Furthermore, after OPA1 mutation, total cellular protein was extracted, and Co-IP was used to detect whether OPA1 interacted with RNF31 in cells. Using the OPA1 antibody as the bait protein, Western blot was performed after IP to detect the expression of OPA1 and RNF31.

Ubiquitination analysis

IP was used to identify the ubiquitination levels of the differential mitochondrial-associated protein OPA1. Cells were treated with dimethyl sulfoxide (DMSO) or 10 μM MG132 for 15 h [33], and the groups were designated as DMSO and MG132. In addition, the following subgroups were included: Control, sh-NC, sh-OTULIN, Control + MG132, sh-NC + MG132, and sh-OTULIN + MG132.

Cycloheximide (CHX) analysis

The CHX assay was used to determine protein half-life by blocking protein translation. After knockdown of OTULIN or RNF31, cells were treated with 50 µg/mL CHX (ab120093, Abcam), and at different time points (0, 2, 4, and 6 h), an equal number of cells was collected to analyze the stability of the differential mitochondrial-associated protein OPA1 [33].

Molecular docking

ClusPro software was used for docking analysis between OPA1 and RNF31, employing a semi-empirical free energy force field to predict the binding affinity between receptor and ligand. Subsequently, PyMOL was used for visualization analysis, indicating that the complex could stably bind within the protein cavity and interact with surrounding amino acids.

Statistical analysis

Data analysis was conducted with GraphPad Prism 8.0. Results are expressed as the mean ± standard deviation. The Kolmogorov–Smirnov test and exploratory descriptive statistics were applied to evaluate whether the data met the assumptions of normality and homogeneity of variance. Data that conformed to normal distribution and homogeneity of variance were analyzed using parametric tests. Specifically, Student’s t-test was conducted to compare two groups. For comparisons involving multiple groups, one-way analysis of variance (ANOVA) was performed, followed by Tukey’s post hoc test. Additionally, data comparisons between groups at different time points were analyzed using two-way ANOVA with Bonferroni post hoc testing. P < 0.05 was considered statistically significant.

Results

Neonatal mice exposed to hyperoxia exhibited lung injury accompanied by dysregulation of deubiquitinating enzyme OTULIN expression

To clarify the effects of hyperoxia exposure at different time points on lung tissue structure and OTULIN expression in neonatal mice, newborn mice were exposed to air (21% O₂) or hyperoxia (95% O₂) from birth until P7 or P14. Alveolar structural development was evaluated by HE staining, and OTULIN protein expression levels were detected by Western blot. HE staining results showed that the lung tissue of mice in the Air group exhibited a normal alveolar structure, with thin and intact alveolar walls, uniform alveolar cavity size, and well-formed secondary crests. In contrast, the lung tissue of mice in the Hyperoxia (P7) group showed obvious alveolar simplification, characterized by enlarged alveolar cavities, thinning and rupture of alveolar walls, and reduced secondary crests. The lung structure injury in the Hyperoxia (P14) group was more significant, with further expansion and fusion of alveolar cavities and severe damage to the alveolar wall structure, presenting typical pathological changes of arrested alveolarization. These findings suggest that hyperoxia-induced lung injury exhibits a time-dependent aggravation pattern (Supplementary Fig. S1A). Morphometric analysis showed that, compared with the Air group, the MLI was significantly increased and RAC was significantly decreased in the Hyperoxia (P7) group. With the extension of hyperoxia exposure to P14, MLI was further significantly increased, while RAC continued to decrease. These results indicate that hyperoxia exposure leads to increased MLI and decreased RAC, and the degree of injury is aggravated with the extension of exposure time (Supplementary Fig. S1B). Western blot analysis showed that the expression level of OTULIN protein was significantly upregulated in a time-dependent manner after hyperoxia exposure. Compared with the Air group, OTULIN expression in lung tissue was significantly increased in the Hyperoxia (P7) group, and the OTULIN expression level in the Hyperoxia (P14) group was further significantly higher than that in the Hyperoxia (P7) group, suggesting that OTULIN may be involved in regulating hyperoxia-induced lung injury in neonatal mice (Supplementary Fig. S1C). These results indicated that hyperoxia exposure not only caused impaired alveolar structural development in neonatal mice (manifested as arrested alveolarization with increased MLI and decreased RAC) but also significantly upregulated the expression of OTULIN in lung tissue, and both changes exhibited time-dependent characteristics, suggesting that OTULIN may play an important role in the occurrence and development of hyperoxic lung injury. Furthermore, BPD model was established by exposing newborn mice to hyperoxia (95% O₂) from birth until day 14, with 21% O₂ serving as the control. HE staining showed that, compared with the Air group, lung tissue in the Hyperoxia group exhibited disorganized structure, with thickened septa, enlarged and fused alveoli, reduced alveolar number, and uneven alveolar size, consistent with the classic pathological features of BPD [34], confirming successful model establishment (Fig. 1A). Quantitative analysis of lung morphology showed that, compared with the Air group, the Hyperoxia group exhibited higher MLI and lower RAC, further indicating the destructive effect of hyperoxia on alveolarization (Fig. 1A). Masson staining revealed a significant increase in fibrosis in the lung tissue of the Hyperoxia group compared with the Air group (Fig. 1B).

Fig. 1.

Neonatal mice exposed to hyperoxia exhibited lung injury accompanied by dysregulation of deubiquitinating enzyme OTULIN expression. Newborn mice delivered on the same day (within 12 h of birth) were pooled and divided into two groups: the Air group and the Hyperoxia group. A Hematoxylin–eosin staining of the morphological structure of mouse lung tissue and assessment of mean linear intercept and radial alveolar count. B Masson staining showing fibrosis in mouse lung tissue. C Western blot analysis of OTUD1, PSMD14, USP14, and OTULIN expression in lung tissue, with β-actin as the loading control. D. Immunohistochemical analysis of OTULIN expression in lung tissue. Scale bars: 100 μm (100 ×) and 25 μm (400 ×). *P < 0.05, ***P < 0.001. n = 6 biologically independent samples. Results are expressed as the mean ± standard deviation. The Kolmogorov–Smirnov test and exploratory descriptive statistics were used to assess whether the data met the assumptions of normality and homogeneity of variance. The data conformed to normal distribution and homogeneity of variance and were therefore analyzed using parametric tests. Specifically, Student’s t-test was used for comparisons between two groups

Western blot was then performed to detect the expression of ubiquitination-related proteins OTUD1, PSMD14, USP14, and OTULIN. Compared with the Air group, PSMD14 and USP14 expression levels in lung tissue of the Hyperoxia group were increased, whereas no significant difference in OTUD1 expression was observed (Fig. 1C). In addition, OTULIN was highly expressed in the lung tissue of the Hyperoxia group compared with the Air group (Fig. 1C), with OTULIN showing the most pronounced increase. Therefore, we focused on the expression and underlying mechanism of the deubiquitinating enzyme OTULIN in BPD. IHC staining further demonstrated that OTULIN expression was highly expressed in the lung tissue of mice in the Hyperoxia group compared with the Air group (Fig. 1D). These results indicate that hyperoxia exposure induces lung injury and upregulation of deubiquitinating enzyme OTULIN expression in neonatal mice.

Hyperoxia-induced neonatal mouse lung tissue exhibited disrupted mitochondrial homeostasis

Research has shown that mitochondrial dysfunction occurs in neonatal BPD [9]. Therefore, we further analyzed mitochondrial quality control-related indicators in hyperoxia-induced mouse lung injury. GO enrichment analysis suggested that OTULIN may be associated with mitochondrial organization, mitochondrial transport, mitophagy, mitochondrial disassembly, mitochondrial fission, and mitochondrial fusion (Fig. 2A). TOM20 is a mitochondrial marker protein located on the outer mitochondrial membrane [35]. Western blot and IF results showed that hyperoxia exposure decreased TOM20 protein expression in the lung tissue of neonatal mice (Fig. 2B, C). Flow cytometry further confirmed that hyperoxia exposure led to a decrease in mitochondrial membrane potential and an increase in ROS levels in the lung tissue of neonatal mice (Fig. 2D, E). Next, we assessed the expression of proteins related to mitochondrial dynamics, including fusion, fission, and autophagy, by Western blot. The results showed that hyperoxia exposure upregulated the expression of OPA1, PINK1, PARK2, BNIP3L, and LC3 II/LC3 I, downregulated the expression of BNIP3 and FUNDC1, and did not significantly alter p-DRP1 Ser-616/DRP1 levels (Fig. 2F). These results indicate that hyperoxia exposure disrupted mitochondrial homeostasis in the lung tissue of neonatal mice, characterized by a reduction in mitochondrial number and increased levels of mitochondrial fusion and autophagy.

Fig. 2.

Hyperoxia-induced neonatal mouse lung tissue exhibited disrupted mitochondrial homeostasis. A The GSE25286 dataset was downloaded from the GEO database and converted into an expression matrix, and the corresponding group information was obtained. R was used to process the expression matrix and analyze differential expression. OTULIN-related genes were screened by Pearson correlation analysis (P < 0.05, |cor|> 0.5), and 2982 genes were obtained. According to the screening results from the previous step, the clusterProfiler R package was used to perform Gene Ontology enrichment analysis. Newborn mice delivered on the same day (within 12 h of birth) were mixed and divided into two groups: the Air group and the Hyperoxia group. B Western blot analysis of TOM20 expression in lung tissue, with β-actin as the loading control. C Immunofluorescence staining of TOM20 expression in lung tissue. TOM20 is shown in green. Scale bar: 25 μm (400 ×). D Flow cytometric detection of mitochondrial membrane potential in lung tissue. E Flow cytometric analysis of mitochondrial ROS levels in lung tissue. F Western blot analysis of proteins related to mitochondrial fusion (OPA1), mitochondrial fission (DRP1 and p-DRP1 Ser-616), and mitophagy (PINK1, PARK2, BNIP3L, BNIP3, FUNDC1, and LC3) in lung tissue, with β-actin as the loading control. ***P < 0.001. n = 6 biologically independent samples. Results are expressed as the mean ± standard deviation. The Kolmogorov–Smirnov test and exploratory descriptive statistics were used to assess whether the data met the assumptions of normality and homogeneity of variance. The data conformed to normal distribution and homogeneity of variance and were therefore analyzed using parametric tests. Specifically, Student’s t-test was used for comparisons between two groups

Overexpression of OTULIN promoted the maintenance of mitochondrial homeostasis in alveolar epithelial cells after hyperoxia exposure

The aforementioned animal experiments demonstrated that hyperoxia exposure led to upregulation of OTULIN expression and disruption of mitochondrial homeostasis in lung tissue. Subsequently, we conducted cellular experiments to further investigate the impact of modulating OTULIN expression on mitochondrial homeostasis in pulmonary epithelial cells. qRT-PCR and Western blot results showed significant changes in OTULIN expression, confirming the effectiveness of overexpression and knockdown (Supplementary Fig. S2A).

Next, we assessed the expression of proteins related to mitochondrial dynamics (fusion, fission, and autophagy) in human alveolar epithelial A549 cells and mouse alveolar epithelial MLE-12 cells by Western blot. Compared with the Control group, the Hyp group showed reduced expression of BNIP3, TOM20, and FUNDC1, and increased expression of OPA1, PINK1, PARK2, BNIP3L, p-DRP1 Ser-616/DRP1, and LC3 II/LC3 I. After OTULIN knockdown, TOM20 expression decreased, whereas the expression of OPA1, p-DRP1 Ser-616/DRP1, PINK1, PARK2, BNIP3L, and LC3 II/LC3 I further increased, with no significant changes in BNIP3 and FUNDC1 expression. After OTULIN overexpression, TOM20 expression increased, whereas the expression of OPA1, p-DRP1 Ser-616/DRP1, PINK1, PARK2, BNIP3L, and LC3 II/LC3 I decreased, with no significant changes in BNIP3 and FUNDC1 expression. Among these, the changes in OPA1, PINK1, PARK2, and BNIP3L expression were the most pronounced (Fig. 3A). In addition, compared with the Control group, the Hyp group showed decreased mitochondrial membrane potential and increased ROS levels in A549 and MLE-12 cells. After knocking down OTULIN, the mitochondrial membrane potential further decreased, whereas ROS levels further increased. In contrast, after overexpressing OTULIN, the mitochondrial membrane potential in A549 and MLE-12 cells increased, and ROS levels decreased (Fig. 3B, C). IF further showed that, compared with the Control group, TOM20 expression was decreased in the Hyp group. After OTULIN knockdown, TOM20 expression further decreased, whereas OTULIN overexpression increased TOM20 expression (Fig. 3D). These results suggested that overexpression of OTULIN promoted the maintenance of mitochondrial homeostasis in alveolar epithelial cells after hyperoxia exposure.

Fig. 3.

Overexpression of OTULIN promoted the maintenance of mitochondrial homeostasis in alveolar epithelial cells after hyperoxia exposure. OTULIN was knocked down or overexpressed, and A549 and MLE-12 cells were transfected with sh-OTULIN or overexpression (oe)-OTULIN plasmids for 48 h before hyperoxia exposure. The groups were designated as follows: Control, Hyp, Hyp + sh-NC, Hyp + sh-OTULIN, Hyp + oe-NC, and Hyp + oe-OTULIN. A Western blot analysis of mitochondrial protein TOM20, the mitochondrial fusion-related protein OPA1, mitochondrial fission-related proteins (DRP1 and p-DRP1 Ser-616), and mitophagy-related proteins (PINK1, PARK2, BNIP3L, BNIP3, FUNDC1, and LC3), with β-actin as the loading control. B JC-1 staining of mitochondrial membrane potential in cells. Mitochondrial membrane potential was indicated by the ratio of red to green fluorescence. A reduction in the red/green fluorescence ratio reflects loss of mitochondrial membrane potential. C MitoSOX fluorescence staining of mitochondrial ROS in cells. D Immunofluorescence staining of TOM20 expression. TOM20 is shown in green. *P < 0.05, **P < 0.01, ***P < 0.001. n = 3 biologically independent samples. Results are expressed as the mean ± standard deviation. The Kolmogorov–Smirnov test and exploratory descriptive statistics were used to assess whether the data met the assumptions of normality and homogeneity of variance. The data conformed to normal distribution and homogeneity of variance and were therefore analyzed using parametric tests. For comparisons among multiple groups, one-way analysis of variance (ANOVA) was performed, followed by Tukey’s post hoc test

To verify the in vivo relevance of the in vitro experimental results, we isolated primary lung epithelial cells from mouse lung tissue. IF staining of KRT8 is an effective tool for identifying mouse primary lung epithelial cells and studying their function and status during lung injury and repair processes [36]. By identifying KRT8-positive expression (Supplementary Fig. S2B), we confirmed the successful isolation of primary mouse lung epithelial cells. This provided a reliable cell model for subsequent experiments. Then we conducted OTULIN knockdown experiments in primary mouse lung epithelial cells. As shown in Supplementary Fig. S2C, compared with the sh-NC group, OTULIN expression was significantly reduced in the sh-OTULIN#1 and sh-OTULIN#2 groups, confirming effective knockdown. Furthermore, after knocking down OTULIN, we treated primary mouse lung epithelial cells with Hyperoxia and detected the expression levels of proteins related to mitochondrial dynamics (fusion, fission, and autophagy). Compared with the Control + sh-NC group, the Hypoxia + sh-NC group showed significantly reduced mRNA and protein expression levels of BNIP3, FUNDC1, and TOM20, and significantly increased mRNA and protein expression levels of BNIP3L, PARK2, OPA1, and PINK1. After further OTULIN knockdown, TOM20 expression was decreased, whereas BNIP3L and PARK2 expression were increased. The protein expression levels of OPA1 and PINK1 were also increased, although no significant changes were observed at the mRNA level. No significant changes were observed in the mRNA or protein expression levels of BNIP3 and FUNDC1 (Supplementary Fig. S2D, E). These results indicate that OTULIN knockdown exacerbates mitochondrial damage in primary mouse lung epithelial cells following hyperoxia exposure, which is consistent with findings in A549 and MLE-12 cells and suggests similar underlying mechanisms.

OTULIN regulated the ubiquitination levels of differential mitochondrial-associated proteins

Next, we further explored the intrinsic mechanisms by which OTULIN regulates differential mitochondrial-associated proteins. First, IF double staining analysis was used to examine the co-localization of OTULIN with differential mitochondrial-associated proteins (OPA1, PINK1, PARK2, and BNIP3L; Fig. 4A). Co-IP was then used to identify and screen differential mitochondrial-associated proteins that directly interacted with OTULIN. Among these proteins, OTULIN interacted with PINK1 and OPA1, whereas no interaction was detected with PARK2 or BNIP3L (Fig. 4B). Since OTULIN inhibited the ubiquitination level of PINK1, and PINK1 maintained protein stability through LUBAC [20], we further focused on the mechanism of interaction between OTULIN and OPA1. IP analysis showed that the differential mitochondrial-associated protein OPA1 could undergo ubiquitination (Fig. 4C). Furthermore, IP analysis showed that OTULIN knockdown increased the ubiquitination level of OPA1 (Fig. 4D). Additionally, CHX analysis was used to assess the stability of OPA1. As time progressed, OPA1 protein expression gradually decreased, and the degradation of OPA1 protein in the sh-OTULIN group was lower than that in the sh-NC group (Fig. 4E). These results indicate that OTULIN regulates the ubiquitination levels of differential mitochondrial-associated proteins.

Fig. 4.

OTULIN regulated the ubiquitination levels of differential mitochondrial-associated proteins. A Immunofluorescence staining analysis of the co-localization of OTULIN with differential mitochondrial-associated proteins (OPA1, PINK1, PARK2, and BNIP3L). B Co-IP identification and screening of differential mitochondrial-associated proteins (OPA1, PINK1, PARK2, and BNIP3L) that directly interact with OTULIN. C, D IP analysis of the ubiquitination levels of the differential mitochondrial-associated protein OPA1. E After knocking down OTULIN, cells were treated with 50 µg/mL CHX (ab120093, Abcam), and at different time points (0, 2, 4, and 6 h), an equal number of cells was collected to analyze the stability of the differential mitochondrial-associated protein OPA1. ***P < 0.001. n = 3 biologically independent samples. Results are expressed as the mean ± standard deviation. The Kolmogorov–Smirnov test and exploratory descriptive statistics were applied to evaluate whether the data met the assumptions of normality and homogeneity of variance. The data conformed to normal distribution and homogeneity of variance and were therefore analyzed using parametric tests. Comparisons between groups at different time points were analyzed by two-way analysis of variance with Bonferroni post hoc test

E3 ubiquitin ligase RNF31 bound to the mitochondrial function-related protein OPA1 and promoted its protein stability

The above-mentioned results identified OPA1 as a differential mitochondrial protein that interacted with OTULIN and was affected by its deubiquitinating activity. To further explore the ubiquitination mechanism of OPA1, we used the online software UbiBrowser and found that OPA1 interacted with two subunits of the ubiquitin E3 ligase complex LUBAC, namely RNF31 and RBCK1. LUBAC is composed of three subunits: the core ubiquitin ligase RNF31/HOIP, the regulatory protein RBCK1/HOIL1, and SHARPIN [37]. The ubiquitin E3 ligase complex LUBAC is usually associated with the deubiquitinating enzyme OTULIN and regulates the linear ubiquitination status of substrates [33]. First, Co-IP results showed that OPA1 interacted with RNF31, whereas no interaction with RBCK1 was detected (Fig. 5A). Therefore, we subsequently focused on RNF31. Next, we knocked down RNF31 and observed a reduction in the expression levels of both RNF31 and OPA1 (Fig. 5B). CHX analysis of mitochondrial-related protein stability showed that, as time progressed, OPA1 protein expression decreased. The stability of the OPA1 protein in the sh-RNF31 group was lower than that in the sh-NC group (Fig. 5C). OTULIN exerts an antagonistic effect on the E3 ligase LUBAC [38]. Therefore, after identifying RNF31 as the E3 ubiquitin ligase that interacted with the mitochondrial protein OPA1, we further analyzed the impact of OTULIN modulation on RNF31 expression and on the binding between OPA1 and RNF31. Western blot analysis revealed that knocking down OTULIN increased the expression of RNF31 and OPA1, while overexpressing OTULIN decreased the expression of RNF31 and OPA1 (Fig. 5D). On the basis of the UbiBrowser prediction of the RNF31 recognition domain in OPA1 (291–469), we further determined the binding region experimentally using OPA1–wt (full length: 1–960), OPA1-mut#1 (1–469/Δ 470–960), and OPA1-mut#2 (1–290/Δ 291–960). Co-IP identified that the binding region through which RNF31 directly interacted with OPA1 was located within amino acids 291–468 (Fig. 5E). Molecular docking further confirmed the binding between OPA1 and RNF31 (Fig. 5F). These results indicated that E3 ubiquitin ligase RNF31 bound to the mitochondrial function-related protein OPA1 and promoted its protein stability.

Fig. 5.

E3 ubiquitin ligase RNF31 bound to the mitochondrial function-related protein OPA1 and promoted its protein stability. A Co-IP identification and screening of the E3 ubiquitin ligases RNF31 and RBCK1 that directly interacted with the differential mitochondrial protein OPA1. RNF31 was knocked down, and the groups were designated as sh-NC and sh-RNF31. B Western blot analysis of the expression of mitochondrial protein OPA1 and E3 ubiquitin ligase RNF31, with β-actin as the loading control. C After RNF31 knockdown, cells were treated with 50 µg/mL CHX (ab120093, Abcam), and at different time points (0, 2, 4, and 6 h), an equal number of cells was collected to analyze the stability of the differential mitochondrial-associated protein OPA1. D Western blot analysis of the expression of RNF31 and OTULIN, with β-actin as the loading control. OPA1 was mutated, and the groups were designated as follows: OPA1–wt (full length: 1–960), OPA1-mut#1 (1–469/Δ 470–960), and OPA1-mut#2 (1–290/Δ 291–960). E Co-IP identification of the direct interaction between E3 ubiquitin ligase RNF31 and the differential mitochondrial protein OPA1. F Molecular docking analysis of the binding between mitochondrial protein OPA1 and RNF31. *P < 0.05, **P < 0.01, ***P < 0.001. n = 3 biologically independent samples. Results are expressed as the mean ± standard deviation. The Kolmogorov–Smirnov test and exploratory descriptive statistics were used to assess whether the data met the assumptions of normality and homogeneity of variance. The data conformed to normal distribution and homogeneity of variance and were therefore analyzed using parametric tests. Specifically, Student’s t-test was used for comparisons between two groups. For comparisons among multiple groups, one-way analysis of variance (ANOVA) was performed, followed by Tukey’s post hoc test. In addition, comparisons between groups at different time points were analyzed by two-way analysis of variance with Bonferroni post hoc test

OTULIN/E3 ubiquitin ligase RNF31 targeted mitochondrial protein OPA1 to suppress hyperoxia-induced mitochondrial dysfunction

We further validated the mechanism involving OTULIN/E3 ubiquitin ligase RNF31 and the mitochondrial protein OPA1 through cellular experiments. Compared with the Control group, OPA1 expression was increased in the Hyp group. After knocking down OTULIN, OPA1 expression further increased. After subsequent RNF31 knockdown, OPA1 expression decreased (Fig. 6A–C). In addition, compared with the Control group, mitochondrial membrane potential was reduced and ROS levels were increased in the Hyp group. Following OTULIN knockdown, mitochondrial membrane potential further decreased, whereas ROS levels further increased. In contrast, after knocking down RNF31, mitochondrial membrane potential increased and ROS levels decreased (Fig. 6D–G). These results suggest that OTULIN/E3 ubiquitin ligase RNF31 targets the mitochondrial protein OPA1 to suppress hyperoxia-induced mitochondrial dysfunction.

Fig. 6.

OTULIN/E3 ubiquitin ligase RNF31 targeted mitochondrial protein OPA1 to suppress hyperoxia-induced mitochondrial dysfunction. OTULIN and RNF31 were knocked down, and A549 and MLE-12 cells were transfected with sh-OTULIN and sh-RNF31 for 48 h before hyperoxia exposure. The groups were designated as follows: Control, Hyp, Hyp + sh-NC, Hyp + sh-OTULIN, Hyp + sh-OTULIN + sh-NC, and Hyp + sh-OTULIN + sh-RNF31. A Western blot analysis of the expression of the mitochondrial fusion-related protein OPA1, with β-actin as the loading control. B, C Immunofluorescence staining of OPA1 expression. D, E JC-1 staining of mitochondrial membrane potential in cells. Mitochondrial membrane potential was indicated by the ratio of red to green fluorescence. A reduction in the red/green fluorescence ratio reflects loss of mitochondrial membrane potential. F, G MitoSOX fluorescence staining of mitochondrial ROS in cells. *P < 0.05, **P < 0.01, ***P < 0.001. n = 3 biologically independent samples. Results are expressed as the mean ± standard deviation. The Kolmogorov–Smirnov test and exploratory descriptive statistics were used to assess whether the data met the assumptions of normality and homogeneity of variance. The data conformed to normal distribution and homogeneity of variance and were therefore analyzed using parametric tests. For comparisons among multiple groups, one-way analysis of variance (ANOVA) was performed, followed by Tukey’s post hoc test

Knockdown of OTULIN exacerbated hyperoxia-induced lung injury in mice, while overexpression of OTULIN alleviated hyperoxia-induced lung injury

Finally, we investigated the effects of OTULIN knockdown and overexpression on hyperoxia-induced neonatal lung injury in vivo. HE staining revealed that, compared with the Air group, lung tissue structure in the Hyperoxia group was disorganized. OTULIN knockdown exacerbated hyperoxia-induced lung tissue injury in neonatal mice (Fig. 7A), whereas OTULIN overexpression alleviated lung injury (Supplementary Fig. S3A).

Fig. 7.

Knockdown of OTULIN exacerbated hyperoxia-induced lung injury in mice. Newborn mice delivered on the same day (within 12 h of birth) were mixed and divided into two groups: the Air group and the Hyperoxia group. Mice in the Hyperoxia + shRNA negative control (sh-NC) and Hyperoxia + sh-OTULIN groups were then subjected to hyperoxia exposure and received intraperitoneal injections of 30 μL sh-NC/sh-OTULIN lentivirus (1 × 10⁸ TU/mL) every 3 days. A, B Hematoxylin–eosin staining showing the morphological structure of mouse lung tissue, together with assessment of mean linear intercept and radial alveolar count. C qRT-PCR analysis of OTULIN expression in lung tissue. D Western blot analysis of OTULIN, OPA1, TOM20, and RNF31 expression in lung tissue, with β-actin as the loading control. E Flow cytometry detection of mitochondrial membrane potential in lung tissue. F Flow cytometry analysis of mitochondrial ROS levels in lung tissue. *P < 0.05, ***P < 0.001. n = 6 biologically independent samples. Results are expressed as the mean ± standard deviation. The Kolmogorov–Smirnov test and exploratory descriptive statistics were used to assess whether the data met the assumptions of normality and homogeneity of variance. The data conformed to normal distribution and homogeneity of variance and were therefore analyzed using parametric tests. For comparisons among multiple groups, one-way analysis of variance (ANOVA) was performed, followed by Tukey’s post hoc test

Quantitative analysis of lung morphology showed that, compared with the Air group, mice in the Hyperoxia group exhibited higher MLI and lower RAC. Following OTULIN knockdown, MLI further increased and RAC further decreased (Fig. 7B). In contrast, after OTULIN overexpression, MLI decreased and RAC increased (Supplementary Fig. S3B). Additionally, compared with the Air group, OPA1 and OTULIN were highly expressed in lung tissue of the Hyperoxia group, whereas TOM20 and RNF31 expression levels were decreased. After OTULIN knockdown, OTULIN and TOM20 expression levels were reduced, whereas OPA1 and RNF31 expression levels were increased (Fig. 7C, D). Conversely, after OTULIN overexpression, OTULIN and TOM20 expression levels were increased, whereas OPA1 and RNF31 expression levels were reduced (Supplementary Fig. S3C, D).

Compared with the Air group, mitochondrial membrane potential was reduced and ROS levels were increased in the Hyperoxia group. After OTULIN knockdown, mitochondrial membrane potential further decreased and ROS levels further increased (Fig. 7E, F). In contrast, after OTULIN overexpression, mitochondrial membrane potential increased and ROS levels decreased (Supplementary Fig. S3E, F). These results indicate that OTULIN knockdown exacerbates hyperoxia-induced lung injury in neonatal mice, whereas OTULIN overexpression alleviates hyperoxia-induced lung injury.

Discussion

BPD is the most common serious complication in extremely preterm infants and infants with very low birth weight [39]. Despite advances in neonatal care, the incidence of BPD remains high among extremely preterm infants [40]. However, the pathogenesis of BPD has not been fully elucidated and warrants further investigation. In this study, we explored the mechanisms of the deubiquitinating enzyme OTULIN/E3 ubiquitin ligase in hyperoxia-induced lung injury through in vivo and in vitro experiments. We found that hyperoxia exposure upregulated the expression of the deubiquitinase OTULIN, while downregulating the expression of the E3 ubiquitin ligase RNF31. This synergistic effect reduced the stability of the mitochondrial protein OPA1 by mediating its linear ubiquitination, thereby attenuating hyperoxia-induced lung injury in neonatal mice. These findings also provide preliminary insight into the roles of deubiquitinating enzyme OTULIN/E3 ubiquitin ligase RNF31, and mitochondrial protein OPA1 in BPD.

The mechanisms by which hyperoxia contributes to BPD are multifactorial and involve oxidative stress, alveolar simplification, pulmonary vascular abnormalities, and immune responses, among other processes. Research has shown that a hyperoxic environment can lead to mitochondrial dysfunction in alveolar epithelial cells, thereby affecting cellular metabolic function, reducing oxidative phosphorylation and ATP production, and ultimately leading to alveolar simplification [41]. Additionally, type 2 innate lymphoid cells have been shown to modulate hyperoxia-induced lung injury by regulating Th17 cell responses [42]. In hyperoxia-induced BPD models, nintedanib has been reported to preserve alveolar and vascular growth, improve lung function, and prevent pulmonary arterial hypertension [43]. These studies provide a deeper understanding of the mechanisms by which hyperoxia contributes to BPD and offer new strategies and potential pharmacological targets for the treatment of BPD.

Deubiquitinating enzymes play a vital role in regulating protein stability and function within cells [44]. In BPD research, the role of deubiquitinating enzymes is gradually attracting attention. Oxidative stress is an important factor in BPD, and deubiquitinating enzymes may participate in the occurrence and development of BPD by affecting oxidative stress responses. For example, SIRT1 activity may be affected by oxidative stress, thereby influencing alveolarization and cell apoptosis [45]. In this study, we found through in vivo experiments that hyperoxia exposure not only caused impaired alveolar structural development in neonatal mice (manifested as arrested alveolarization with increased MLI and decreased RAC) but also significantly upregulated the expression of OTULIN in lung tissue, and both changes exhibited a time-dependent manner. Our findings suggest that the deubiquitinating enzyme OTULIN may play a significant role in the occurrence and development of BPD, potentially by regulating protein stability and influencing lung development and repair processes.

The morphology and distribution of mitochondria are regulated by the dynamic balance between mitochondrial fusion and fission. A hyperoxic environment may disrupt this balance, leading to abnormal mitochondrial morphology and impaired mitochondrial function and quality control [46]. Under hyperoxic conditions, mitochondria produce increased levels of ROS, which can damage mitochondrial DNA (mtDNA) and lead to mitochondrial dysfunction [47]. Mitophagy is the process by which damaged or dysfunctional mitochondria are selectively removed. A hyperoxic environment may affect the efficiency of mitophagy, leading to the accumulation of damaged mitochondria and consequently affecting cellular homeostasis [48]. In this study, we confirmed through animal experiments that mitochondrial homeostasis was disrupted in the lung tissue of neonatal mice exposed to hyperoxia. Notably, we observed that both hyperoxia-induced upregulation of OTULIN and downregulation of RNF31 could reduce OPA1 protein stability by mediating linear ubiquitination, thereby exerting a protective effect against hyperoxia-induced lung injury. Therefore, we next sought to explore how OTULIN upregulation interacts with RNF31 downregulation in the pathogenesis of BPD.

OTULIN is a deubiquitinating enzyme that plays a key role in regulating the levels of linear ubiquitin chains (Met1-linked polyubiquitin chains) [49]. Research has shown that, by controlling Met1-linked ubiquitination in the nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) signaling pathway, OTULIN may affect inflammation-mediated lung injury and repair processes [50], which are key components of the pathogenesis of BPD. OTULIN is also involved in regulating cellular autophagy, including mitophagy, thereby helping to clear damaged mitochondria, maintain mitochondrial quality control, and support cellular energy metabolism and responses to oxidative stress [51]. Additionally, OTULIN participates in type I interferon signaling and antiviral responses by regulating the Met1 ubiquitination of signal transducer and activator of transcription 1 (STAT1) [52]. These findings reveal the multiple roles of OTULIN in maintaining mitochondrial homeostasis and function within cells, as well as its complex role in regulating cell fate. However, there are no direct studies investigating OTULIN in the context of BPD.

In this study, through further mechanistic investigation, we found that OTULIN overexpression promoted the maintenance of mitochondrial homeostasis in alveolar epithelial cells following hyperoxic exposure, supporting the notion that OTULIN upregulation serves as a protective adaptive response. Alveolar type II epithelial cells, such as MLE-12 cells, have significant repair and regenerative capacity following lung injury, and they can differentiate into alveolar type I epithelial cells to restore alveolar gas exchange function [53]. Although A549 cells are a cancer cell line, their origin and characteristics allow them to mimic certain functions of alveolar type II epithelial cells, especially in terms of cell proliferation and differentiation [54]. A549 and MLE-12 cells exhibit biological properties similar to those of alveolar type II epithelial cells in vitro. For example, A549 cells can secrete the surfactant protein SP-A [54], whereas MLE-12 cells are more similar to primary alveolar type II epithelial cells [55]. These cells help maintain the stability and integrity of alveolar structure by regulating mitochondrial dynamics, thereby playing important roles in alveolar epithelial repair and regeneration, as well as in immune defense [56]. The mitochondrial homeostasis regulatory mechanisms exhibited by these cells in vitro are similar to those by which alveolar type II epithelial cells maintain alveolar structure and function in vivo. Therefore, we have reason to believe that the mitochondrial homeostasis regulatory mechanisms observed in vitro may also be relevant in vivo and may have important effects on pulmonary metabolism and inflammatory responses. Furthermore, OTULIN regulated the ubiquitination levels of differential mitochondrial-associated proteins. Our research represents a preliminary exploration of the potential mechanisms of OTULIN in BPD.

OTULIN specifically hydrolyzes Met1-linked ubiquitin chains conjugated by LUBAC [57]. LUBAC is an E3 ubiquitin ligase complex composed of three subunits: RNF31/HOIP, RBCK1/HOIL1, and SHARPIN [37]. LUBAC catalyzes the formation of linear ubiquitin chains within cells, a process that is crucial for various cellular activities, including inflammatory responses, cell death, and immune responses [58]. It has been reported that LUBAC participates in the regulation of the NF-κB signaling pathway by catalyzing the formation of linear ubiquitin chains [59]. RNF31 is a key factor in initiating mitophagy and promoting mitochondrial homeostasis, and reduced RNF31 expression is associated with decreased mitochondrial membrane potential and impaired mitophagy [60]. OPA1 is a protein primarily localized to the mitochondrial inner membrane, where it plays a crucial role in regulating mitochondrial morphology and function, especially during mitochondrial fusion and fission [61]. Studies have shown that OPA1 influences cellular energy metabolism and cell death pathways by regulating mitochondrial morphology and function [62]. These studies highlight the important role of LUBAC in cellular signal transduction and disease pathogenesis, as well as the role of OPA1 in mitochondrial function and cell death. In this study, we confirmed that E3 ubiquitin ligase RNF31 bound to the mitochondrial function-related protein OPA1 and promoted its protein stability. Furthermore, OTULIN/E3 ubiquitin ligase RNF31 targeted the mitochondrial protein OPA1 to inhibit hyperoxia-induced mitochondrial dysfunction. Finally, through in vivo experiments, we demonstrated that knockdown of OTULIN exacerbated hyperoxia-induced lung injury in neonatal mice, whereas overexpression of OTULIN alleviated hyperoxia-induced lung injury. These findings suggest that OTULIN upregulation represents a compensatory protective mechanism in response to mitochondrial dysfunction. Our research represents a preliminary exploration of the mechanism involving OTULIN/E3 ubiquitin ligase RNF31 and the mitochondrial protein OPA1 in hyperoxia-induced lung injury in neonatal mice.

However, this study still has several limitations. First, sex was not used as a stratification factor during randomization, and the current sample size was underpowered to detect potential sex-specific differences in hyperoxia-induced mitochondrial phenotypes. Future studies with larger, sex-balanced cohorts are warranted to determine whether the OTULIN–OPA1 axis operates similarly in male and female neonates. Furthermore, we plan to establish a new hyperoxia model in future studies, with a preset sample size of n = 15 for each sex, to detect a 30% difference with 90% power and to systematically assess the effect of sex on pulmonary mitochondrial homeostasis. Second, this study mainly focused on the OTULIN/RNF31–OPA1 axis, whereas mitochondrial quality control involves a multilayered regulatory network. For example, the phosphorylation level of DRP1 was not significantly affected by hyperoxia, suggesting that the fission pathway may not be the primary mechanism; however, this requires further verification in combination with dynamic morphological observations. Thirdly, the observation that hyperoxia downregulates RNF31 while upregulating OTULIN suggests a complex regulatory interplay that may involve feedback mechanisms or cell-type specific responses. Future studies should investigate the upstream signaling pathways that regulate RNF31 and OTULIN expression under hyperoxic conditions. In addition, the protective effect of RNF31 inhibition has thus far only been demonstrated in cell models. Therefore, animal studies using conditional RNF31knockdown or selective inhibitors are required to verify whether targeting RNF31 can indeed attenuate lung injury in the hyperoxic BPD model. Although the interaction domain between OPA1 and RNF31 has been identified, the structural basis of this interaction and the impact of posttranslational modifications, such as phosphorylation, on their binding have not yet been elucidated. Finally, given that OTULIN upregulation appears to be a protective response, therapeutic strategies aimed at enhancing OTULIN activity or expression may represent a promising approach for BPD treatment, warranting further investigation in preclinical models. In the future, human organoids or clinical samples should be used to verify the conservation of this pathway, and the interaction of OTULIN/RNF31 between alveolar epithelial cells and immune cells should also be explored to more fully clarify its pathological significance in BPD.

Conclusions

In this study, we assessed the mechanism of action of the deubiquitinating enzyme OTULIN in BPD and determined that hyperoxia exposure upregulates OTULIN expression while downregulating RNF31 expression in neonatal lung tissue. The potential mechanism of action was related to the mitochondrial protein OPA1 and mitochondrial homeostasis. These findings provide a new perspective for understanding the molecular mechanisms of BPD. In future studies, we will further elucidate the precise mechanisms involving OTULIN/E3 ubiquitin ligase RNF31 and OPA1 within cells, as well as their potential applications in disease treatment.

Abbreviations

BPD

Bronchopulmonary dysplasia

Hyp

Hyperoxia

HE

Hematoxylin–eosin

MLI

Mean linear intercept

RAC

Radial alveolar count

qRT-PCR

Quantitative real-time polymerase chain reaction

IF

Immunofluorescence

IHC

Immunohistochemistry

IOD

Integrated optical density

ROS

Reactive oxygen species

GO

Gene Ontology

Co-IP

Co-immunoprecipitation

CHX

Cycloheximide

mtDNA

Mitochondrial DNA

Author contributions

LH: conceptualization, formal analysis, visualization, writing—original draft; QL: formal analysis, investigation, visualization, writing—original draft; AZ, YL, FH, JW, MT, and DW: data curation, formal analysis, visualization; MZ: resources, supervision, writing—review and editing; XW: methodology, project administration, writing—review and editing. All authors reviewed the results and approved the final version of the manuscript.

Funding

This study was supported by National Natural Science Foundation of China (82200017), The Science and Technology Innovation Program of Hunan Province, China (2023RC3198), Natural Science Foundation of Hunan Province, China (2025JJ50622, 2025JJ90191, 2026JJ30097, 2026JJ81762), Scientific Research Project of Hunan Provincial Health Commission, China (20255112), Hunan Provincial Department of Education Scientific Research Project (no. 24B0069, 25A0327), and the Huxiang High-Level Talent Agrregation Project for 2021-Innovative Talent (no. 2021RC5014).

Availability of data and materials

Data will be made available on request.

Declarations

Ethics approval and consent to participate

The study was approved by The Animal Ethics Committee of Hunan Provincial People’s Hospital (2024, no. 140) on 29 July 2024. The animal experiments were complied with ICLAS Ethical Guidelines and relevant laws and regulations.

Consent for publication

Not applicable.

Competing interests

The authors have no conflicts of interest to declare.