Bone Marrow Stromal Cell-Secreted Extracellular Vesicles Containing miR-34c-5p Alleviate Lung Injury and Inflammation in Bronchopulmonary Dysplasia Through Promotion of PTEN Degradation by Targeting OTUD3

ABSTRACT Background: Bronchopulmonary dysplasia (BPD) is the predominant chronic disorder in preterm neonates. This study explored impacts of miR-34c-5p carried by bone marrow stromal cells-secreted extracellular vesicles (BMSC-EVs) on BPD progression.Methods: A BPD mouse model was established, followed by measurement of miR-34c-5p, OTUD3, and PTEN expression. EVs were isolated from BMSCs transfected with miR-34c-5p mimic or mimic NC and intratracheally injected into mice. CD31 and Ki67 expression was detected and the pathological changes of lung tissues and lung function indexes were observed for mice. A neonatal human pulmonary microvascular endothelial cell (HPMEC) model was developed with hyperoxia, followed by co-culture with extracted EVs and ectopic experiments for measurement of cell viability, migration, and angiogenesis. IL-4, IL-13, IL-1β, and IL-6 levels were measured in cell supernatants and lung tissues. Dual-luciferase reporter, ubiquitination, Co-IP, and RIP assays were adopted to determine the relationship among miR-34c-5p, OTUD3, and PTEN.Results: Lung tissues of BPD mice had downregulated miR-34c-5p expression and upregulated OTUD3 and PTEN expression. BMSC-EVs and BMSC-EVs-miR-34c-5p treatment improved lung injury and alveolar structure, decreased lung resistance and IL-4, IL-13, IL-1β, and IL-6 levels, and elevated dynamic lung compliance in BPD mice, as well as enhanced proliferation, angiogenesis, and migration and restrained inflammation in HPMECs. Mechanistically, miR-34c-5p negatively targeted OTUD3 which restrained ubiquitination to promote PTEN protein stabilization. Upregulation of OTUD3 or PTEN negated the changes in the proliferation, angiogenesis, migration, and inflammation of hyperoxia-treated HPMECs induced by BMSC-EVs-miR-34c-5p.Conclusion: BMSC-EVs-miR-34c-5p alleviated lung injury and inflammation in hyperoxia-induced BPD by blocking the OTUD3/PTEN axis. GRAPHICAL ABSTRACT


Introduction
As a respiratory disorder, bronchopulmonary dysplasia (BPD) occurs in preterm newborns and results in chronic respiratory issues (Principi et al. 2018). The risk factors for BPD include low birth weight, prematurity, chorioamnionitis, intrauterine growth restriction, smoking, ethnicity or race, and heredity (Thebaud et al. 2019). Additionally, inflammation is also a crucial contributor to BPD (Savani 2018). The main "old" features of BPD were hyperinflated, scarred, and fibrotic lungs caused by lung injury after oxygen toxicity and mechanical ventilation, which have changed to lung growth arrest-caused impaired lung angiogenesis and alveolar simplification with the development of neonatal medicine (Schmidt et al. 2022). Adults with BPD are proven to be severely affected throughout their lives, as they are observed to suffer from aberrant lung functions, decreased exercise tolerance, and a possibly elevated risk of chronic obstructive pulmonary disease (COPD) (Gilfillan et al. 2021). Consequently, searching for novel biomarkers and clarification of molecular mechanisms involved in BPD progression is necessary for the advancement of BPD treatment.
Bone marrow stromal cells (BMSCs) have been documented to alleviate vascular and parenchymal injury in BPD through their paracrine function (Aslam et al. 2009). The paracrine role of MSCs is achieved through the secretion of soluble factors and extracellular vesicles (EVs) which carry proteins, microRNAs (miRs/miRNAs), and mRNAs from parent cells to recipient cells (Keshtkar et al. 2018). As reported, the administration of MSC-EVs has substantial therapeutic promise for the treatment of BPD through the enhancement of macrophage polarization, improvement of angiogenesis and alveolarization, and reduction of collagen density (Guo et al. 2021). Moreover, BMSC-derived exosomes ameliorate lipopolysaccharide (LPS)-induced acute lung injury by delivering miR-384-5p . As reported, abnormal expression of miRNAs is implicated in BPD progression (Ameis et al. 2017). Of note, miR-34c-5p was reported to be downregulated in patients with COPD and might serve as a potential biomarker for early detection of COPD (Akbas et al. 2012). Moreover, Dong et al. reported low miR-34c-5p expression in a BPD mouse model through a comprehensive analysis (Dong and Zhang 2021). Therefore, miR-34c-5p may be involved in the effects of BMSC-EVs on BPD progression.
The Starbase database utilized in our study predicted that ovarian tumor deubiquitinase 3 (OTUD3) bound to miR-34c-5p. OTUD3, a deubiquitinase, can stabilize proteins by deubiquitinating K48-linked ubiquitination and modulate protein activity by deubiquitinating K11-or K63-linked ubiquitination (Shen et al. 2021). Reportedly, OTUD3 suppressed tumorigenesis by promoting the stability of phosphatase and tensin homolog (PTEN) through deubiquitination (Yuan et al. 2015). Moreover, PTEN downregulation participates in the repressive impacts of microvesicles from human umbilical cord MSCs on alveolarization and lung inflammation in BPD rats (Zhou et al. 2022). These backgrounds contribute to the hypothesis that BMSC-EVs containing miR-34c-5p (BMSC-EVs-miR -34c-5p) might affect BPD progression through the OTUD3/PTEN axis. Therefore, this study was designed to confirm this speculation, providing a novel therapeutic direction for BPD.

miR-34c-5p expression was low in BPD mice
To delve into the mechanism of BPD onset, neonatal mice were exposed to hyperoxia from P0 to P14 to establish a BPD mouse model. The results of lung function indexes showed that neonatal mice in the BPD group had remarkably higher LR and lower Cydn as compared to mice in the RA group (Figure 1a-b).
As demonstrated in Figure 1c, HE staining results revealed that in the lung tissues collected at P14, the alveolar structure was normal and alveoli were uniform and numerous in the RA group. The BPD group showed obvious disorganized lung tissue structure with simplified and enlarged alveoli, indicating impaired alveolar development. In addition, alveolar development was quantified through statistical analysis of the number and diameter of alveoli and the cross-sectional area of the alveolar cavity in lung histopathological sections. The results presented a dramatically reduced RAC value and increased the diameter of the alveoli and the cross-sectional area of the alveolar cavity in neonatal mice in the BPD group versus the RA group (Figure 1d-f). ELISA results indicated the elevated level of IL-1β, IL-6, IL-4, and IL-13 in the mouse lung tissues in the BPD group ( Figure 1g). qRT-PCR data also evidenced that mice in the BPD group had notably augmented IL-1β and IL-6 levels in lung tissues (Figure 1h). Immunohistochemistry showed noticeably lower CD31 (a marker of lung endothelial angiogenesis) and Ki67 (a marker of proliferation) positive rates in mouse lung tissues in the BPD group than in the RA group (Figure 1i-j). Moreover, miR-34c-5p expression was observed to be appreciably poorer in the BPD group than in the RA group by qRT-PCR detection ( Figure 1k).

BMSC-EVs-miR-34c-5p ameliorated hyperoxia-induced lung injury in BPD mice
Next, we investigated the role of stem cells overexpressing miR-34c-5p on the hyperoxiainduced BPD mouse model and HPMEC model. First, BMSCs as well as extracted BMSC-EVs were identified. The results displayed the successful isolation and extraction of BMSC-Figure 1. BPD mice have low miR-34c-5p expression. (a-b) Detection and analysis of lung function indexes lung resistance (a) and dynamic lung compliance (b); (c) HE staining to observe lung histopathology; (d-f) statistical analysis of alveolar number (d), alveolar diameter (e), and cross-sectional area of alveolar cavity (f) on lung histopathology sections; (g) ELISA to determine IL-1β, IL-6, IL-4, and IL-13 levels in lysates of mouse lung tissues; (h) qRT-PCR to determine IL-1β and IL-6 levels in mouse lung tissues; (i-J) immunohistochemistry to examine the expression of CD31 (i) and Ki67 (j); (k) qRT-PCR to determine miR-34c-5p expression in mouse lung tissues. n = 6 mice/group. The measurement data were stated as mean ± standard deviation. The t-test was adopted for comparisons of mean values of samples between two groups. * indicates P < 0.05 compared to the RA group. Figure S1). To discuss the alleviatory effects of BMSC-EVs on the hyperoxia-induced BPD mouse model, it was first necessary to determine whether BMSC-EVs were internalized by lung tissues. Accordingly, BMSC-EVs were evidenced to be internalized by lung tissues of BPD mice (Supplementary Figure S2A).

EVs (Supplementary
To examine the repressive impacts of BMSC-EVs and BMSC-EVs-miR-34c-5p on the hyperoxia-induced BPD mouse model, neonatal mice were exposed to hyperoxia immediately after birth until P14 and injected with BMSC-EVs (40 μL) or BMSC-EVs-miR-34c-5p (20 μg of exosomal protein) intratracheally at P4 and P8, with PBS as the control. At P14, mice were euthanized and specimens were harvested. qRT-PCR results displayed that miR-34c-5p expression was elevated in the mouse lung tissues of the EVs (BPD + BMSC-EVs) group relative to the PBS (BPD + PBS) group and further elevated in the mouse lung tissues of the EVs-miR-34c-5p (BPD + BMSC-EVs-miR-34c-5p) group versus the EVs group ( Figure 2a).
As displayed in Figure 2(b-c), the alterations in LR and Cydn values were reversed in the EVs and EVs-miR-34c-5p groups as compared to the PBS group, and better improvement was observed in the EVs-miR-34c-5p group, indicating that BMSC-derived EVs-miR-34c-5p treatment improved lung injury in hyperoxia-induced BPD mice.
As revealed by HE staining, the lung histopathological structure and alveolar simplification were improved in the EVs group versus the PBS group and were further improved in the EVs-miR-34c-5p group versus the EVs group ( Figure 2d). The EVs group had considerably elevated RAC values and decreased the diameter of alveoli and the cross-sectional area of the alveolar cavity, with further improvement in the EVs-miR-34c-5p group versus the EVs group (Figure 2e-g). ELISA results indicated that IL-1β, IL-6, IL-4, and IL-13 levels were substantially lower in mouse lung tissues of the BMSC-EVs and BMSC-EVs-miR-34c-5p groups than in the PBS group ( Figure 2h). qRT-PCR assays also presented lower IL-1β and IL-6 levels in mouse lung tissues of the BMSC-EVs and BMSC-EVs-miR-34c-5p groups ( Figure 2i).
Immunohistochemistry results demonstrated a marked enhancement in CD31 and Ki67 positive rates in mouse lung tissues of the BMSC-EVs and BMSC-EVs-miR-34c-5p groups in contrast to the PBS group (Figure 2j-k). In summary, BMSC-derived EVs-miR-34c-5p ameliorated hyperoxia-induced lung injury and vascular impairment and suppressed inflammation.

BMSC-EVs-miR-34c-5p enhanced the proliferative, migratory, and angiogenic capacities of HPMECs and restrained inflammation
Afterward, we further examined the action mechanism of BMSC-EVs-miR-34c-5p in the amelioration of lung injury in BPD mice. Firstly, it was verified that BMSC-EVs could be internalized by HPMECs (Supplementary Figure S2B). Subsequently, BMSCs were respectively transfected with miR-34c-5p mimic and miR-NC, followed by isolation of EVs. qRT-PCR assays presented that the transfection was successful and that miR-34c-5p mimic transfection in BMSCs augmented the content of miR-34c-5p in the isolated EVs (Supplementary Figure S2c-d).
Next, an in vitro cell model was constructed in HPMECs with hyperoxia. BMSC-EVs-miR-NC or BMSC-EVs-miR-34c-5p were co-cultured with hyperoxia-stimulated HPMECs. qRT-PCR results illustrated that miR-34c-5p expression was elevated in hyperoxia- Figure 2. BMSC-EVs-miR-34c-5p improves lung injury in hyperoxia-induced BPD mice. (a) qRT-PCR detection of miR-34c-5p expression in mouse lung tissues; (b-c) measurement of lung function indexes lung resistance (b) and dynamic lung compliance (c); (d) HE staining for observing lung histopathology; (e-g) statistical analysis of alveolar number (e), alveolar diameter (f), and cross-sectional area of alveolar cavity (g) on lung histopathology sections; (h) ELISA to test IL-1β, IL-6, IL-4, and IL-13 levels in lysates of mouse lung tissues; (i) qRT-PCR to examine IL-1β and IL-6 levels in mouse lung tissues; (j-k) immunohistochemistry to assess the expression of CD31 (j) and Ki67 (k). n = 6 mice/group. The measurement data were stated as mean ± standard deviation. One-way analysis of variance was employed for comparisons among multiple groups with Tukey's test for post hoc multiple comparisons. #Indicates P < 0.05 compared to the PBS group; & indicates P < 0.05 compared to the EVs group. stimulated HPMECs after co-culture with EVs-miR-NC and further augmented after coculture with EVs-miR-34c-5p ( Figure 3a). The results of CCK-8, scratch, and tubule formation assays exhibited that the proliferative, migratory, and angiogenic capacities were pronouncedly decreased in the HYP group relative to the control group, higher in the HYP + EVs-miR-NC group than in the HYP group, and increased in the HYP + EVs-miR-34c-5p group compared to the HYP + EVs-miR-NC group (Figure 3b-d). ELISA results indicated that IL-1β and IL-6 levels in the cell supernatants were evidently enhanced in the HYP group as compared to the control group, diminished in the HYP + EVs-miR-NC group versus the HYP group, and reduced in the HYP + EVs-miR-34c-5p group in contrast to the HYP + EVs-miR-NC group (Figure 3e). In conclusion, BMSC-EVs-miR-34c-5p facilitated the proliferation, angiogenesis, and migration of HPMECs and suppressed inflammation in vitro.

miR-34c-5p negatively targeted OTUD3
To further explore the regulatory mechanism of BMSC-EVs-miR-34c-5p on lung injury in BPD mice, the Starbase database was employed for prediction, which showed that OTUD3 was a direct target of miR-34c-5p ( Figure 4a). The dual-luciferase reporter assay was conducted to verify their relationship, which displayed that the relative luciferase activity of cells co-transfected WT-OTUD3 reporter gene vectors and miR-34c-5p mimic was appreciably lower than that of cells co-transfected with mimic NC (P < 0.01). Nevertheless, there was no pronounced difference in the relative luciferase activities of the MUT-OTUD3 reporter gene vector between the miR-34c-5p mimic and mimic NC groups (P > 0.05, Figure 4b). The RIP assay demonstrated that miR-34c-5p and OTUD3 were dramatically enriched in Ago2 immunoprecipitation in contrast to IgG immunoprecipitation ( Figure 4c).
qRT-PCR and western analyses manifested that OTUD3 expression was appreciably higher in the mouse lung tissues of the BPD group than in the RA group and notably lower in the mouse lung tissues of the EVs and EVs-miR-34c-5p groups, especially the EVs-miR -34c-5p group, than in the PBS group (Figure 4d-e). In HPMECs, OTUD3 expression was strikingly elevated in the HYP group versus the control group, decreased in the HYP + EVs-miR-NC group relative to the HYP group, and reduced in the HYP + EVs-miR-34c-5p group compared with the HYP + EVs-miR-NC group (Figure 4f-g). These findings revealed that miR-34c-5p inversely targeted OTUD3.

OTUD3 promoted PTEN protein stabilization through deubiquitination
qRT-PCR and western analyses indicated that PTEN expression in the lung tissues of BPD mice and HPMECs was consistent with the trend of OTUD3 expression (Figure 5a-d).
HPMECs were transfected with oe-OTUD3 to ascertain whether OTUD3 manipulated the stability of PTEN in BPD. As reflected by the results of qRT-PCR and western analyses, oe-OTUD3 did not affect PTEN mRNA expression (Figure 5e) but augmented PTEN protein expression ( Figure 5f). Co-IP results demonstrated that the OTUD3 antibodies aggregated PTEN proteins and that the PTEN antibodies aggregated OTUD3 proteins ( Figure 5g). Western analysis to test ubiquitination levels displayed that oe-OTUD3 strongly restrained the ubiquitination of PTEN in the presence of MG132 (a potent inhibitor of the 26S proteasome) (Figure 5h). To sum up, OTUD3 facilitated PTEN protein stabilization through deubiquitination.

Establishment of a BPD mouse model
Neonatal FVB mice and maternal mice were acquired from Vital River Laboratories (Beijing, China). From birth (P0) to postnatal day 14 (P14), neonatal FVB mice were exposed to room air (RA) or a polymethyl methacrylate chamber containing 85% oxygen, which was ventilated and maintained at 85% oxygen concentration with an oxygen concentration control device (Biospherix, Lacona, NY, USA). Ammonia was filtered with an air purifier and activated carbon. Normoxic and hyperoxic maternal mice were replaced every 48 h to prevent maternal mice from oxygen toxicity (Ahn et al. 2018;You et al. 2020).
Neonatal mice were arranged into five groups: RA (6 mice), BPD (6 mice treated with hyperoxia), EVs (6 mice treated with hyperoxia and BMSC-EVs), EVs-miR-34c-5p (6 mice treated with hyperoxia and BMSC-EVs-miR-34c-5p), and phosphate-buffered saline (PBS, 6 mice treated with hyperoxia and PBS) groups. On day 4 (P4) and day 8 (P8), mice in the EVs and . BMSC-EVs-miR-34c-5p facilitates the proliferation, migration, and angiogenesis of HPMECs and suppresses inflammation. HPMECs were stimulated with hyperoxia to construct an in vitro cell model. (a) qRT-PCR for miR-34c-5p expression in hyperoxia-stimulated HPMECs; (b) CCK-8 assay for the proliferation of HPMECs; (c) scratch assay for the migration of HPMECs; (d) tubule formation assay for the angiogenesis of HPMECs; (e) ELISA assay for IL-1β and IL-6 levels in the cell supernatants. Cell experiments were conducted in three replicates. The measurement data were stated as mean ± standard deviation. Oneway analysis of variance was employed for comparisons among multiple groups with Tukey's test for post hoc multiple comparisons. *indicates P < 0.05 compared to the control group; # indicates P < 0.05 compared to the HYP group; & indicates P < 0.05 compared to the HYP + EVs-miR-NC group. MiR-34c-5p targets and negatively regulates OTUD3. (a) Starbase database to predict the binding sites between OTUD3 and miR-34c-5p; (b) dual-luciferase reporter assay to verify the binding of miR-34c-5p to OTUD3; (c) RIP assay to confirm the binding of miR-34c-5p to OTUD3; (d-e) qRT-PCR (d) and western analysis (e) to assess OTUD3 levels in lung tissues of BPD mice; (f-g) qRT-PCR (f) and western analysis (g) to measure OTUD3 levels in HPMECs. Cell experiments were conducted in three replicates. n = 6 mice/group in animal experimentations. The measurement data were stated as mean ± standard deviation. One-way analysis of variance was employed for comparisons among multiple groups with Tukey's test for post hoc multiple comparisons. *indicates P < 0.05 compared to the mimic NC, IgG, RA, or control group; # indicates P < 0.05 compared to the PBS or HYP group; & indicates P < 0.05 compared to the EVs or HYP + EVs-miR-NC group.
EVs-miR-34c-5p groups were intratracheally injected with 40 µL BMSC-EVs and BMSC-EVs-miR-34c-5p (20 µg exosomal proteins), respectively, while mice in the PBS group were intratracheally injected with 40 µL PBS as a control. After 14 days, mice were euthanized by anesthesia, and lung tissues were attained for histological experiments and immunohistochemistry. Animal experiments gained ratification from the Ethics Committee of the Second Nanning People's Hospital for animal care and use. (e) qRT-PCR to assess OTUD3 and PTEN mRNA levels; (f) western analysis to determine OTUD3 and PTEN protein levels; (g) Co-IP assay to verify the interaction between OTUD3 and PTEN; (h) ubiquitination assay to examine the ubiquitination level of PTEN. Cell experiments were conducted in three replicates. n = 6 mice/group in animal experimentations. The measurement data were stated as mean ± standard deviation. The t-test was adopted for comparisons of mean values of samples between two groups, and one-way analysis of variance was utilized for comparisons among multiple groups, with Tukey's test for post hoc multiple comparisons. *Indicates P < 0.05 compared to the RA, control, or oe-NC group; # indicates P < 0.05 compared to the PBS or HYP group; & indicates P < 0.05 compared to the EVs or HYP + EVs-miR-NC group.

Lung function test
Lung resistance (LR) and dynamic lung compliance (Cydn) were examined in neonatal mice to determine the impact of altered lung structure on lung function. Mice were anesthetized with sodium pentobarbital (40 mg/kg), intubated after tracheostomy, and mechanically ventilated using a computer-controlled small animal ventilator (EMKA Technologies, Falls Church, VA, USA) with a tidal volume of 8 mL/kg, a positive endexpiratory pressure (PEEP) of 3 cm H 2 O, and a respiratory rate of 150 breaths/min. LR and Cydn were measured every 5 s. Figure 6. BMSC-EVs-miR-34c-5p stimulates proliferation, migration, and angiogenesis and curbs inflammation in HPMECs through the OTUD3/PTEN axis activation. Hyperoxia-stimulated HPMECs were transfected with oe-OTUD3 or oe-PTEN. (a) qRT-PCR to determine OTUD3 and PTEN mRNA expression; (b) western analysis to measure OTUD3 and PTEN protein expression; (c) CCK-8 assay to test the proliferation of HPMECs; (d) scratch assay to evaluate the migration of HPMECs; (e) tubule formation assay to examine the angiogenesis of HPMECs; (f) ELISA to assess IL-1β and IL-6 levels in the cell supernatants. Cell experiments were conducted in three replicates. The measurement data were stated as mean ± standard deviation. The t-test was adopted for comparisons of mean values of samples between two groups, and one-way analysis of variance was utilized for comparisons among multiple groups, with Tukey's test for post hoc multiple comparisons. *Indicates P < 0.05 compared to the oe-NC or EVs-miR-34c-5p + oe-NC group.

Histological examination of lung tissues
Lung tissues were embedded with paraffin and serially sectioned for HE staining. Three random sections of each specimen were observed for morphological changes in lung tissues under a microscope, with five random fields. Radial alveolar counts (RAC) were counted in each field with Image-Pro Plus 6.0 software, with the diameter of the alveoli and the cross-sectional area of the alveolar cavity recorded and averaged.

Immunohistochemistry
Subsequent to 48-h fixation in 4% paraformaldehyde, the lung tissues were made into paraffin sections. The sections were baked for 20 min, followed by conventional xylene dewaxing and washing with distilled water. Subsequent to three washes with PBS, the sections were left to stand with 3% H 2 O 2 for 10 min at room temperature. Thereafter, the sections were subjected to three washes with PBS and antigenic thermal repair, followed by three washes with PBS. The sections were sealed with normal goat serum sealing solution for 20 min at room temperature, and then the excess liquid was shaken off. Subsequently, the sections were subjected to overnight incubation with primary antibodies (Cell Signaling Technologies, Beverly, MA, USA; 1:50) against Ki67 (#12202) and CD31 (#77699) at 4°C. Then, the sections were washed with PBS and probed with secondary antibodies at room temperature for 2 h. After three washes with PBS, the sections were developed with diaminobenzidine for 1-3 min, and then the development was terminated. The sections were subjected to 3-min nucleus staining with hematoxylin before dehydration, transparency, sealing, and observation and image acquisition under the microscope.

Isolation, culture, and identification of BMSCs
Primary BMSCs were isolated from the femur and tibia of FVB mice (5-7 weeks old), centrifuged, concentrated, and plated. BMSCs were cultivated in Dulbecco's Modified Eagle Medium/F12 encompassing 1% penicillin/streptomycin and 10% fetal bovine serum with the conditioned medium renewed every 2-3 days. After three passages, cells that reached approximately 80% confluence were chosen for the follow-up experimentations.
BMSCs at passage 3 were cultured for 21-day osteogenic induction or chondrogenic induction (Sigma-Aldrich, Burlington, MA, USA) or 14-day adipogenic induction (Cyagen, Guangzhou, China) and respectively stained with alizarin red, alcian blue, and oil red O, followed by photography. Moreover, flow cytometry was employed to examine the expression of stem cell negative markers CD34 and CD45 and positive markers CD29 and CD90.

Isolation of EVs derived from BMSCs
When cell confluence reached 80%, BMSCs at passages 3-5 were rinsed twice with PBS and further cultured with serum-free medium for 48 h. The supernatants were collected and centrifuged at 300 g for 10 min to remove the cells. Next, the supernatants were obtained and centrifuged at 2000 g for 10 min to remove the dead cells. Subsequently, the supernatants were collected and centrifuged at 10,000 g for 10 min to remove the cell debris. The supernatants were attained for 70 min of centrifugation at 100,000 g on an ultracentrifuge, followed by the discarding of the supernatant. The precipitates were then washed with PBS and centrifuged again at 100,000 g for 70 min to remove impurity proteins and harvest EV precipitates. Finally, the EV precipitates were resuspended with PBS. After the estimation of the concentration, the EV precipitates were stored at −80°C.

Identification of EVs from BMSCs
EVs (20 μL) were added to the copper woven mesh. Afterward, EVs were dried with filter paper after 1 min and then added with 1 drop of 1% uranyl acetate. Subsequent to 1 min, EVs were dried with filter paper. The morphology of EVs was viewed with transmission electron microscopy (TEM, Hitachi, Tokyo, Japan). The expression of the EV markers CD81 and CD63, the endoplasmic reticulum marker Calnexin, and the Golgi matrix protein GM130 was determined with western analysis.

Nanoparticle tracking analysis (NTA)
EVs (10 μL) were diluted in 1 mL filtered PBS, and size distribution was determined with a Zetasizer Nano S90 (Malvern Panalytical, Malvern, UK). Video images of particle motion were captured and analyzed with NTA software.

Cell culture and processing
Neonatal human pulmonary microvascular endothelial cells (HPMECs) were obtained from ScienCell Research Laboratories (San Diego, CA, USA) and cultured at 37°C with endothelial cell medium (ScienCell Research Laboratories) in a 5% CO 2 incubator.

Cell transfection and grouping
When 80-90% confluence was reached, BMSCs were transfected with miR-34c-5p or miRnegative control (NC) following the protocols of Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) and classified into miR-34c-5p and miR-NC groups. The miR-34c-5p mimic and mimic NC were purchased from RiboBio (Guangdong, China) and transfected at a concentration of 100 nM. After transfection, EVs were isolated by ultracentrifugation and respectively noted as the EVs-miR-34c-5p group and the EVs-miR-NC group.
To probe whether OTUD3 modulated PTEN stability to function in BPD, HPMECs were transfected with overexpression (oe)-OTUD3 or oe-NC and assigned into oe-OTUD3 and oe-NC groups.
HPMECs were co-cultured with 80 µg/mL PKH-26-labeled EVs for 0 h and 24 h, fixed with 4% paraformaldehyde for 20 min, stained with DAPI for 15 min, and finally observed with the fluorescence microscope.

Enzyme-linked immunosorbent assay (ELISA)
Corresponding ELISA kits were applied to examine IL-1β (R&D Systems, Minneapolis, MN, USA), IL-6 (R&D Systems), IL-4 (R&D Systems), and IL-6 (Abcam, Cambridge, UK) levels in the lysate of mouse lung tissues and the supernatant of cells. All operations were conducted strictly in accordance with the manuals of the ELISA kits.

Cell counting kit (CCK)-8 assay
The viability of HPMECs was assessed with CCK-8 kits (ab228554, Abcam) as instructed in the manuals of the manufacturer.

Scratch assay
Cells were spread in six-well plates. When cells grew to 90% confluence, straight lines were drawn in the plates with the tip of a 200 μL pipette. The floating cells were washed with PBS and the medium was replaced with a serum-free medium. Next, the blank gap between the cells was observed with a low-magnification phase contrast microscope (Olympus, Tokyo, Japan). The changes of blank gaps were observed after 24-h culture. The migratory rate = (scratch distance at 0 h -scratch distance at 24 h)/scratch distance at 0 h.

Tubule formation assay
Cells (8 × 10 4 cells/well) were plated on 48-well plates coated with Matrigel (356230, Corning, Tewksbury, MA, USA) deprived of growth factors to assess angiogenesis. Each group was set up with three wells. Five h later, images of tubule formation in at least three random fields were randomly captured with a light microscope and analyzed with Image J, with tube length and the number of branches as the evaluation criteria.

Quantitative reverse transcription polymerase chain reaction (qRT-PCR)
Total RNA of tissues, cells, and EVs was extracted and reversely transcribed in the light of the manuals of TaqMan Reverse Transcription Kits (ABI Inc., Oyster Bay, NY, USA). The PCR primers (Table 1) were synthesized by Sangon (Shanghai, China). Gene expression was tested with a LightCycler 480 (Roche Diagnostics, Indianapolis, IN, USA) fluorescent quantitative PCR instrument, and reactions were conducted following the instructions of the SYBR Green PCR kit (Thermo Fisher, Wilmington, DE, USA). PCR cycling conditions were pre-denaturation at 95°C for 10 min and 40 cycles of denaturation at 95°C for 10 s, annealing at 57°C for 30 s, and extension at 72°C for 20 s. The internal reference of miR-34c-5p was U6, and that of mRNA was glyceraldehyde-3-phosphate dehydrogenase (GAPDH). The threshold value was manually selected at the lowest point of the parallel rise of each logarithmic amplification curve to acquire the cycle threshold (Ct) value for each reaction tube. The 2 −ΔΔCt method was adopted to analyze the data, and 2 −ΔΔCt indicated the ratio of the expression of target genes between the experimental group and the control group.

Dual-luciferase reporter assay
The reporter gene plasmids of wild-type OTUD3 (WT-OTUD3) and mutant OTUD3 (MUT-OTUD3) were constructed and co-transfected with miR-34c-5p mimic or mimic NC into cells. Forty-eight h later, Firefly and Renilla luminescence signals were examined in cells as per the protocols of the dual-luciferase reporter gene assay kit (Beyotime, Shanghai, China) on a luminescence detector (Promega Corporation, Madison, WI, USA). The value of target luminescence/internal reference luminescence was calculated as the relative luciferase activity.

RNA immunoprecipitation (RIP) assay
Magna RIP kits (Millipore, Billerica, MA, USA) were employed for this assay. Lung tissues were lysed with RIP buffer and centrifuged to harvest the lysate supernatants. The supernatants were incubated with anti-argonaute 2 (Ago2) or Immunoglobulin G (IgG) antibodies at 4ºC. After incubation, beads with proteinase K were added to digest proteins. RNA was extracted with Trizol for immunoprecipitation, and the co-precipitated RNA was purified and subjected to qRT-PCR analysis.

Co-immunoprecipitation (Co-IP)
The whole cell lysates derived by centrifugation were subjected to probing with OTUD3 (AA 192-220, Antibodies-online), PTEN (ab267787, Abcam), or normal IgG antibodies and protein G-Agarose beads (Roche Diagnostics Ltd, Shanghai, China) overnight at 4°C. Western analysis was employed to determine the protein expression in the immune complexes.
Lysates were centrifuged to obtain cytoplasmic proteins and incubated overnight with anti-Myc antibodies (ab9106, 1:1000, Abcam). Then, the sample was slowly shaken and incubated for 4 h at 4°C with protein A/G agarose beads (100 μL) to capture antigen-antibody mixtures. After three washes with lysis buffer, the antigen-antibody mixtures were boiled in 2 × SDS loading buffer to release proteins before immunoblotting was conducted with anti-HA antibodies (ab9110, 1:2000, Abcam).

Statistical analysis
Cell experiments were conducted in three replicates. Statistical analysis was conducted with GraphPad Prism 8.0 software. The measurement data were stated as mean ± standard deviation. The t-test was adopted to compare the mean values of samples between two groups, and one-way analysis of variance was carried out for comparisons among multiple groups. Tukey's test was utilized for post hoc multiple comparisons. P < .05 was deemed as a statistically significant difference.

Discussion
BPD mainly happens in infants born with lungs that are still in transition from the canalicular to the saccular phase, and premature delivery interrupts lung growth in these stages, which is often exacerbated by prenatal and postnatal events that can result in the arrest of alveolar development and lung vascular (Hwang and Rehan 2018). Hyperoxia and chronic inflammation are thought to be essential mechanisms of BPD (Wang and Tsao 2020). Although advancements in neonatal care have increased survival and reduced incidence, there has been limited progress in lowering BPD rates. In addition, there are limited treatment options for protecting the delicate developing lungs and strategies for treating lung injury (Tracy and Berkelhamer 2019). Therefore, we pay attention to the biomarkers that influence inflammation and lung injury in BPD and their promising mechanisms. In this study, hyperoxia was employed to induce a BPD mouse model and neonatal HPMEC model to delve into the impacts of miR-34c-5p delivered by BMSC-EVs on inflammation and lung injury in BPD. Consequently, the present study discovered that BMSC-EVs-miR-34c-5p improved lung injury and inflammation in hyperoxia-induced BPD via the inactivation of the OTUD3/PTEN axis. MSCs can be implicated in alleviation of BPD by repressing inflammation and improving angiogenesis and alveolar structure (Bonadies et al. 2020). As a kind of MSCs, BMSCs are a potential therapeutic option to alleviate the severity of BPD because they facilitate lung repair through secretion of paracrine factors or direct regeneration . EVs are membrane structures derived from cells and deliver miRNAs to be involved in the regulation of diseases . As reported, BMSCs-derived exosomes enhance the survival of lung epithelial cells to suppress hyperoxia-induced lung injury (Wu et al. 2021). Similarly, our data disclosed that BMSC-EVs ameliorated lung injury, alveolar structure, and inflammation in BPD mice. More importantly, it has been reported that BMSC-EVs can carry miR-34c-5p to control the progression of several diseases. For instance, BMSCderived exosomal miR-34c-5p improves renal interstitial fibrosis through various proteins (Hu et al. 2022). BMSC-EVs containing miR-34c-5p mitigated edematous lung injury by elevating the expression of the γ-epithelial sodium channel (Hua et al. 2022). miR-34c-5p is specifically associated with human lung disorders in that its expression is elevated in the normal development of lungs (Savarimuthu Francis et al. 2014). For instance, miR-34c-3p downregulation participates in epithelial-mesenchymal transition in lung fibrosis regulated by long non-coding RNA ATB (Xu et al. 2021). Moreover, miR-34c-5p expression was reported to be low in a BPD mouse model (Dong and Zhang 2021). Consistently, the present study identified down-regulated miR-34c-5p expression in the hyperoxia-induced BPD mouse model. However, the specific impact of miR-34c-5p on BPD remains to be discovered. Of note, a prior study unraveled that overexpression of miR-34c-5p increased Cydn and mean alveolar number and suppressed inflammation in COPD rats through CCL22 (Gao et al. 2019). Concordantly, BMSC-EVs-miR-34c-5p treatment further improved lung injury, alveolar structure, and inflammation in BPD mice relative to BMSC-EVs. Likewise, versus BMSC-EVs, BMSC-EVs-miR-34c-5p treatment also further enhanced proliferative, migratory, and angiogenic capacities and restrained inflammation in HPMECs.
miRNAs are a class of small endogenous non-coding RNAs that can modulate the expression of target genes (Kabekkodu et al. 2018). Starbase database prediction showed OTUD3 as a direct target of miR-34c-5p, which was confirmed by dual-luciferase reporter and RIP assays. Ubiquitination, among the most essential post-translational modification types, is a multi-step enzymatic process that participates in variable cell biological activities . In contrast, deubiquitination triggered by deubiquitinases detaches ubiquitin and regulates signaling factor stability (Park et al. 2020). OTUD3, an acetylationdependent deubiquitinase, can stabilize PTEN protein Yuan et al. 2015), which is similar to our result that OTUD3 promoted the protein stabilization of PTEN through deubiquitination. Although OTUD3 expression has been detected in lung cancer (Deng et al. 2019;Zhang et al. 2020), little is known about its expression and role in other lung injuries, especially BPD. Intriguingly, our data revealed that OTUD3 expression was high in the BPD mouse model. PTEN is a dual phosphatase with lipid phosphatase and protein activities and a negative modulator of the phosphatidylinositol-3-kinase/AKT pathway, a primary pathway for cell survival and growth . Reportedly, PTEN expression was visibly elevated in the lungs of LPS-induced BPD rats (Zhou et al. 2022). In this study, PTEN expression was also high in the BPD mouse model. PTEN participates in the arrested alveolar development in neonatal rats by hyperoxia exposure through mitotic impairment (Yu et al. 2020). Moreover, human umbilical cord MSC-derived small EVs mitigated lung injury in a BPD rat model by accelerating cell survival and angiogenesis via PTEN downregulation (You et al. 2020). Considering these references and observations, rescue experiments were conducted in hyperoxia-induced HPMECs. The data elucidated that upregulation of OTUD3 or PTEN reversed the elevation in proliferation, angiogenesis, and migration and the reduction in inflammation of hyperoxia-treated HPMEC caused by BMSC-EVs-miR-34c-5p.
In conclusion, the present study disclosed that the neonatal mice with hyperoxia-induced BPD had lowly expressed miR-34c-5p and highly expressed OTUD3 and PTEN. Intratracheal administration of BMSC-EVs and BMSC-EVs-miR-34c-5p alleviated lung injury and inflammation in hyperoxia-induced BPD via the elevation of PTEN degradation by targeting OTUD3 (Figure 7). The relevance of the present study is to shed light on the impact of BMSC-EVs carrying miR-34c-5p on lung injury and inflammation in BPD, which may indicate BMSC-EVs carrying miR-34c-5p as a novel therapeutic target for preventing or ameliorating clinical BPD. However, the stability and pharmacokinetic and pharmacodynamic properties of BMSC-EVs-miR-34c-5p will be a concern. According to our results and prior research, PTEN downregulation plays a role in the alleviation of BPD. Accordingly, it may be interesting to see the results of direct inhibition of PTEN using small soluble inhibitors in BPD models. However, direct inhibition of PTEN was not performed in this study, because the main purpose of this study was to explore the role and mechanism of BMSC-EVs-miR-34c-5p in BPD. Therefore, further research is warranted.

Acknowledgments
Thanks for the website Biorender.

Ethics statement
Animal experiments gained ratification from the Ethics Committee of the Second Nanning People's Hospital for animal care and use.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
The author(s) reported there is no funding associated with the work featured in this article.

Data availability statement
The datasets used or analyzed during the current study are available from the corresponding author on reasonable request. Figure 7. Graphical abstract. In neonatal mice with hyperoxia-induced BPD, miR-34c-5p was lowly expressed and OTUD3 and PTEN were highly expressed. Intratracheal administration of BMSC-EVs and BMSC-EVs-miR-34c-5p enhanced PTEN degradation by downregulating OTUD3 expression to ameliorate hyperoxia-induced lung injury in mice with BPD.