USP8 promotes the progression of sepsis-induced acute lung injury by interacting with the USP24 to suppress PGC-1α deubiquitination and expression

Abstract

Sepsis is a life-threatening condition with high morbidity and mortality, in which acute lung injury (ALI) represents one of the earliest and most severe complications, yet effective therapeutic strategies remain limited. The molecular mechanisms underlying ALI pathogenesis are still incompletely understood. In this study, we identify ubiquitin-specific protease 8 (USP8) as a critical regulator of mitochondrial homeostasis and ferroptosis during sepsis-induced ALI. Mechanistically, USP8 modulates mitochondrial biogenesis and promotes ferroptosis by interfering with the interaction between USP24 and PGC-1α in a manner independent of its deubiquitinase activity. Notably, genetic ablation of USP8 markedly attenuates ALI; however, this protective effect is strictly dependent on the presence of USP24. Collectively, our findings uncover a previously unrecognized regulatory mechanism by which USP8 controls ferroptotic signaling and highlight USP8 and USP24 as potential therapeutic targets for sepsis-associated acute lung injury.

1. Introduction

Sepsis-induced acute lung injury (ALI) is a manifestation of systemic, infection-driven multi-organ dysfunction, characterized by inflammatory cell infiltration, increased permeability of the alveolar–capillary barrier, and acute diffuse alveolar edema [1]. It is a life-threatening hypoxic respiratory syndrome with a mortality rate approaching 40%. The complexity of pulmonary inflammation, arising from intricate interactions among multiple cell types—including alveolar type II (AT2) epithelial cells—has substantially limited the development of effective therapies for ALI [2]. At present, clinical management of ALI relies largely on supportive strategies such as positive-pressure ventilation, and no disease-modifying treatments are available. Given that AT2 cells possess stem-like properties and can transdifferentiate into alveolar type I cells during lung injury, thereby playing a critical role in maintaining epithelial barrier integrity, we sought to elucidate the molecular pathways governing AT2 cell function in the context of acute lung injury. Defining these mechanisms may enable the identification of rapid and effective therapeutic strategies for ALI and related pathological conditions.

Ubiquitination is a critical post-translational modification (PTM) that regulates diverse cellular processes by controlling protein stability, trafficking, subcellular localization, and protein–protein interactions. A single ubiquitin molecule contains seven lysine residues (K6, K11, K27, K33, K48, and K63), allowing ubiquitin moieties to be assembled into polyubiquitin chains with distinct linkage types, each conferring specific cellular functions. Canonically, K48- or K11-linked polyubiquitin chains target proteins for degradation by the 26S proteasome, whereas K27-linked ubiquitination has been implicated primarily in mitochondrial dynamics and plays an important role in maintaining cellular energy homeostasis. In recent years, mitochondrial dynamics have been extensively investigated across multiple disease models. For example, hematopoietic stem cells enriched in mitochondria exhibit enhanced self-renewal capacity [3], and clinical evidence indicates that mitochondria are essential for maintaining skeletal muscle homeostasis in patients with type 2 diabetes mellitus [4]. Moreover, the role of mitochondria in tumorigenesis and cancer progression has been widely reported [5]. Our previous work demonstrated that preservation of mitochondrial homeostasis in macrophages during the early stages of acute lung injury confers therapeutic benefit [6]. However, the contribution of ubiquitination to mitochondrial quality control in acute lung injury remains largely unexplored.

USP8 is a well-characterized deubiquitinating enzyme that has been extensively studied in non–small cell lung cancer [7]. Recent evidence further indicates that USP8 plays an important role in coordinating ferroptosis and responses to immunotherapy [8]. Here, using a conditional USP8 knockout mouse model, we investigated the previously unrecognized role of USP8 during the development of acute lung injury. We found that, in the presence of USP8, USP8 associates with USP24 and PGC-1α in a manner independent of its deubiquitinating enzymatic activity, thereby limiting mitochondrial biogenesis and promoting ferroptosis. In contrast, in the absence of USP8, USP24 directly interacts with PGC-1α and removes K27-linked ubiquitin chains, leading to increased mitochondrial abundance, suppression of ferroptosis, and restoration of alveolar epithelial barrier integrity. Collectively, our findings delineate a previously unappreciated biological pathway within alveolar type II epithelial cells during acute lung injury and highlight targeting USP8 as a potential therapeutic strategy for restoring lung epithelial barrier function.

2. Materials and methods

2.1. Animals and animals model

All animal experiments were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Animal Care and Use Committee of Renmin Hospital of Wuhan University (approval no. WDRM20220604). Wild-type C57BL/6 mice were purchased from Shulaibao (Wuhan) Biotechnology. USP8^fl/fl and USP24^−/− mice were obtained from Cyagen Biosciences. To achieve lung epithelial cell–specific deletion of USP8, USP8^fl/fl mice were crossed with Sftpc-CreERT2 mice to generate USP8^fl/fl; Sftpc-CreERT2 offspring. For temporal and spatial induction of gene deletion, mice aged 8–10 weeks were administered tamoxifen by oral gavage once every 24 h for five consecutive days, thereby inducing USP8 deletion specifically in lung epithelial cells.

To establish the acute lung injury model, phenol-extracted and purified lipopolysaccharide (LPS; 10 mg/kg) was dissolved in sterile saline and administered via intraperitoneal injection. Acute lung injury was induced 12 h after LPS challenge. Following LPS stimulation, lung tissues were harvested: the left lung was used for histological analysis, and the right lung was snap-frozen in liquid nitrogen and stored at low temperature for subsequent biochemical assays.

2.2. Hematoxylin and eosin (H&E) staining of lung tissue and lung injury scoring

Following LPS challenge, left lung tissues were harvested for histological analysis. Samples were fixed in 4% paraformaldehyde, embedded in paraffin, and sectioned at a thickness of 5 μm. Sections were deparaffinized, rehydrated, and stained with hematoxylin and eosin (H&E), and histopathological changes were examined under a light microscope. Given the variability in lung injury scoring criteria reported in previous studies, we focused on key pathological features, including hyaline membrane formation, neutrophil infiltration, proteinaceous debris, and alveolar septal thickening. These parameters were evaluated to generate a semi-quantitative assessment of lung injury, providing a quantitative measure of the severity of lung damage [9].

2.3. Collection of bronchoalveolar lavage fluid

Upon establishng the acute lung injury model, mice were deeply anesthetized to allow exposure of the trachea through a subtle neck midline incision. A cannulation tube was positioned into the tracheal lumen and anchored with surgical thread. Bronchoalveolar lavage fluid (BALF) was retrieved by gently instilling and slowly aspirating 0.5 mL of sterile saline, repeating the flush cycle three times. The recovered fluid was then processed for subsequent protein and cellular characterization.

2.4. Survival analysis

To further determine the impact of USP8 on survival, USP8^fl/fl; Sftpc-Cre mice and wild-type (WT) littermates were provided with free access to food and water and subjected to the LPS-induced acute lung injury model. Survival was monitored for 7 days following LPS challenge to assess the effect of USP8 deficiency on mortality associated with acute lung injury. Deaths were recorded daily. Survival analysis was performed using the Kaplan–Meier method, and differences between survival curves were evaluated using the log-rank test.

2.5. Isolation of primary murine alveolar type II (AT2) epithelial cells

Primary alveolar type II (AT2) epithelial cells were obtained from USP8^fl/fl; Sftpc-CreERT2 mice and corresponding wild-type littermates. To achieve USP8 ablation, isolated cells were subjected to ex vivo Cre recombinase-mediated recombination. Briefly, after anesthesia, mice were perfused via the right ventricle to minimize blood contamination, and the lungs were subsequently collected for cell isolation. Lung tissues were mechanically minced and enzymatically dissociated using a combination of trypsin and collagenase digestion at 37°C, followed by filtration to obtain a homogeneous cell suspension. AT2 cells were enriched through IgG-mediated negative selection using antibody-coated culture dishes. The purified cells were maintained in F–12K medium containing 10% FBS for 72 h before downstream analyses to facilitate cellular recovery and stabilization.

2.6. Cell lines, transfection, and establishment of stable cell lines

Most in vitro experiments were performed using primary murine AT2 epithelial cells. To further validate our findings, we also employed commonly used lung epithelial cell lines, including the human alveolar epithelial cell line A549 and the HEK293T cell line. A549 cells were cultured in F–12K medium supplemented with 10% FBS, whereas HEK293T cells were maintained in RPMI 1640 medium containing 10% FBS. All cells were cultured at 37°C in a humidified incubator with 5% CO₂.

When cells achieved ∼65% confluency, transient plasmid transfections were executed with Lipofectamine 2000 or polyethyleneimine (PEI) (Beyotime) following the provided protocol, followed by cell harvesting at designated time points for downstream downstream assays. For stable gene silencing or knockout, lentiviral execution vectors (pLKO.1 containing shRNA or sgRNA) were co-introduced into HEK293T packaging cells along with pVSVG and pD8.9 via PEI or Lipofectamine 2000. Virus-containing supernatants were harvested 36–48 h later, filtered, and applied to target cells supplemented with 4 μg/mL polybrene, after which antibiotic selection was applied. To achieve transient overexpression, wild-type or mutated coding regions engineered into pcDNA3.1 were introduced into specified host cells.

2.7. Western blot

Western blot analysis was performed as previously described [10]. Briefly, cells were lysed on ice using RIPA buffer (Servicebio) to extract total protein. For immunoprecipitation experiments, lysates were incubated with magnetic beads conjugated to the appropriate primary antibody to capture target proteins. Proteins were separated by SDS-PAGE at suitable concentrations and then transferred onto PVDF membranes. Membranes were blocked with 5% non-fat milk to reduce nonspecific binding of primary antibodies, followed by incubation with the corresponding secondary antibodies. Protein signals were detected using a chemiluminescence imaging system. Details of the antibodies used in this study are summarized in Supplementary Table 1.

2.8. RNA extraction and quantitative RT-PCR (qRT-PCR)

Total RNA was extracted from tissues or cells using enzyme-free TRIzol reagent. Complementary DNA (cDNA) was synthesized from the isolated RNA using a reverse transcription kit (Vazyme). Quantitative PCR was then performed using SYBR Green chemistry to amplify the cDNA. GAPDH was used as an internal control, and relative gene expression was calculated using control samples as the reference for subsequent analyses.

2.9. Protein stability analysis

To investigate the effect of USP8 on PGC-1α, USP8 expression was knocked down using shRNA, and cycloheximide (CHX) was applied to inhibit protein synthesis. Cells were harvested at indicated time points, and protein levels were analyzed by Western blot using specific antibodies. PGC-1α band intensities were quantified with ImageJ and normalized to the initial time point to assess protein degradation and the impact of USP8 knockdown on protein stability.

2.10. GST pull-down

To produce recombinant N-terminally GST-tagged proteins (wild-type or truncated variants), pGEX-6P-2 vectors were introduced into Escherichia coli. Following expression induction, the resulting GST-fused proteins were captured using GST-agarose magnetic beads. Complex formation was achieved by incubating the purified GST-tagged constructs with the target protein for 3 h at 4°C. After extensive washing to eliminate non-specifically bound components, bound protein complexes were recovered via elution with reduced glutathione buffer.

2.11. RNA sequencing and bioinformatic analysis

Total RNA was extracted from mouse lung tissues using TRIzol reagent according to the manufacturer's instructions. Library preparation and sequencing were performed by Igenebol Biotechnology on the Illumina platform. Sequencing quality was assessed using FastQC, and reads were aligned to the mouse reference genome (GRCm38). Genes with P < 0.05 and |Log₂ fold change| > 1.5 were considered differentially expressed. Differentially expressed genes (DEGs) were subjected to Gene Ontology (GO) and KEGG pathway enrichment analyses using the R package clusterProfiler, and enrichment results were visualized as bubble plots.

2.12. Isolation and transplantation of mitochondria

Mitochondria were isolated using a commercial Mitochondria Isolation Kit (C3602S; Beyotime, Shanghai, China) according to the manufacturer's instructions. Briefly, cells were gently lysed, and functional mitochondria were rapidly separated via selective precipitation and differential centrifugation. For mitochondrial transplantation, the freshly isolated mitochondria were co-cultured with the target cells, allowing internalisation into the recipient cells primarily through endocytosis. To evaluate the intervention effects on endogenous cellular components, the relative abundance and integrity of mitochondrial DNA (mtDNA) were assessed and compared before and after the transplantation.

2.13. Immunoprecipitation–mass spectrometry (IP-MS)

Immunoprecipitation–mass spectrometry (IP-MS) analysis was conducted with the assistance of SpecAlly Life Technology (Wuhan, China). To identify USP8 interactors, AT2 cells expressing empty vector or USP8 were lysed in buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 1% NP-40, 0.1% SDC, 1 mM EDTA, and protease inhibitors) at 4°C for 30 min. Lysates were cleared by centrifugation (12,000 × g, 10 min) and incubated with anti-USP8 antibody overnight at 4°C. Immunoprecipitates were washed with PBS, then denatured, reduced, and alkylated in 1% SDC/100 mM Tris-HCl (pH 8.5) with 10 mM TCEP and 40 mM CAA at 95°C for 10 min. Samples were diluted with water and digested overnight with trypsin (1:50 w/w) at 37°C. Digestion was quenched with TFA, and peptides were desalted using SDB tips.

Peptides were analyzed on an UltiMate 3000 RSLCnano system paired with a Q Exactive HF mass spectrometer (Thermo Fisher). Separation was performed on a C18 analytical column (75 μm × 25 cm, 1.9 μm) at 300 nL/min using water/acetonitrile gradients containing 0.1% formic acid and 3% DMSO. Data were acquired in DIA mode and processed using DIA-NN (v1.8.1) in library-free mode against the Mus musculus UniProt database (2024-09-10). Search parameters included trypsin/P digestion (≤2 missed cleavages), fixed carbamidomethylation (C), and variable oxidation (M) and N-terminal acetylation. MS1/MS2 tolerances were set to 15 ppm with MBR enabled. Protein quantification was normalized via MaxLFQ. Identified interactors are listed in Supplementary Table 2.

2.14. Mitochondrial staining

Mitochondria were stained using MitoTracker (Beyotime) according to the manufacturer's instructions. All staining procedures were performed on live cells. Cells were incubated with MitoTracker for 30 min, after which mitochondrial morphology and distribution were observed using a fluorescence confocal microscope.

2.15. Statistical analysis

Quantitative data are expressed as mean ± SD and were processed using GraphPad Prism (v10.3.0). Two-group evaluations were executed via unpaired, two-tailed Student's t-tests, whereas multi-group datasets were evaluated using one-way ANOVA with Tukey's or Sidak's post hoc comparisons. For survival outcomes, cumulative survival probabilities were estimated using the Kaplan–Meier approach, with inter-group differences evaluated by log-rank test.

3. Results

3.1. Conditional deletion of USP8 in lung epithelium confers protection during sepsis-induced acute lung injury by restoring respiratory function and prolonging survival

To evaluate the role of USP8 in acute lung injury, we first examined its expression in lung tissues from control and LPS-challenged mice and found that USP8 levels were markedly increased following LPS stimulation (Supplementary Fig. 1A). We then generated lung epithelial cell–specific USP8 conditional knockout mice and observed no obvious effects on growth or development before or after USP8 deletion (Fig. 1A and Supplementary Fig. 1B–C). Concurrently, the offspring derived from crosses between USP8^fl/fl and USP8^fl/fl; Sftpc-CreERT2 mice were born at the expected Mendelian ratios, demonstrating normal embryonic survival and inheritance patterns (Supplementary Fig. 1D and E). Analysis of bronchoalveolar lavage fluid (BALF) collected from USP8^fl/fl and USP8^fl/fl; Sftpc-Cre mice before and after LPS challenge revealed that USP8 deletion significantly reduced both total cell counts and protein concentrations in BALF, indicating preservation of alveolar–capillary barrier integrity (Fig. 1B and Supplementary Fig. 1F), which was further supported by increased expression of epithelial barrier–associated genes at the mRNA level (Fig. 1F). Functionally, conditional deletion of USP8 significantly improved survival and restored respiratory function in mice subjected to LPS-induced acute lung injury (Fig. 1C–D). Histological examination by H&E staining showed that USP8 deficiency markedly reduced pulmonary hemorrhage and partially restored alveolar architecture disrupted by LPS exposure (Fig. 1E).

Fig. 1.

Reduced USP8 expression protects against sepsis-induced acute lung injury. (A) Gross morphology of lungs from USP8^fl/fl,USP8^fl/+, and USP8^fl/fl; Sftpc-CreERT2 mice. (B) Protein concentrations in bronchoalveolar lavage fluid (BALF) from USP8^fl/fl and USP8^fl/fl; Sftpc-CreERT2 mice following LPS challenge. (n = 6 mice per group) (C) Kaplan–Meier survival curves of USP8^fl/fl,USP8^fl/+, and USP8^fl/fl; Sftpc-CreERT2 mice subjected to LPS-induced acute lung injury. (n = 20 mice per group) (D) Changes in respiratory function in USP8^fl/fl and USP8^fl/fl; Sftpc-CreERT2 mice before and after LPS stimulation. (n = 6 mice per group) (E) Representative H&E staining of lung sections showing histopathological alterations following LPS challenge. (Scale bar = 50 μm) (F) Quantitative PCR analysis of lung epithelial barrier–related genes in lung tissues after LPS stimulation. (n = 3 mice per group) (G) RNA-seq–based pathway enrichment analysis comparing USP8 conditional knockout mice with control mice after LPS stimulation. (H) Measurement of glutathione (GSH) levels as an indicator of ferroptosis in lung tissues. (n = 6 mice per group) (I) Immunohistochemical staining of 4-hydroxynonenal (4-HNE) in lung sections. (Scale bar = 50 μm).

To further validate the protective effects of USP8 deficiency on pulmonary function, we established another sepsis model utilizing Pseudomonas aeruginosa challenge (Supplementary Fig. 1G–H). Intriguingly, we observed that USP8 knockout consistently exerted a protective effect on lung tissues within this model (Supplementary Fig. 1I and J). Given that sepsis-induced acute lung injury (ALI) primarily progresses through two distinct pathogenic mechanisms—either direct disruption of the alveolar epithelium or indirect tissue injury driven by macrophage activation and subsequent cytokine release—we next sought to elucidate whether macrophages contribute to this USP8-mediated protection. To this end, we generated USP8^fl/fl;LysM-iCre mice to selectively deplete USP8 within the myeloid lineage (Supplementary Fig. 2A–C), which conformed to expected Mendelian inheritance ratios (Supplementary Fig. 2D and E) and exhibited no significant differences in body length or gross organ weights compared to control mice (Supplementary Fig. 2F–I). However, unlike the epithelial knockout, myeloid-specific USP8 depletion failed to yield any significant alterations in alveolar barrier protection (Supplementary Fig. 2J), nor did it produce appreciable changes in overall survival rates or pro-inflammatory cytokine release profiles (Supplementary Fig. 2K and L). Taken together, these findings demonstrate that the protective role of USP8 in preserving lung function relies predominantly on mitigating the disruption of the alveolar epithelial barrier rather than suppressing myeloid-driven inflammatory activation. To explore the underlying mechanisms, we performed RNA sequencing of lung tissues, and KEGG pathway enrichment analysis revealed that differentially expressed genes were predominantly enriched in ferroptosis- and proteasome-related pathways (Fig. 1G). Consistent with these findings, USP8-deficient lungs exhibited restored glutathione (GSH) levels and reduced 4-hydroxynonenal (4-HNE) staining, indicating attenuated lipid peroxidation and ferroptosis following LPS challenge (Fig. 1H–I).

3.2. Knockdown of USP8 restores the expression of the mitochondrial biogenesis gene PGC-1α in alveolar type II epithelial cells

To further investigate the mechanism by which USP8 regulates ferroptosis, we employed multiple ferroptosis inducers to mimic the effects of LPS stimulation (Fig. 2A and Supplementary Fig. 3A). We found that USP8 knockout markedly reduced Leflunomide-induced ferroptosis in AT2 cells (Fig. 2B), which was further supported by changes in intracellular GSH levels and MDA content (Fig. 2C). Notably, inhibition of USP8 significantly reduced cell death and improved cell viability induced by Leflunomide treatment (Fig. 2D and E). Furthermore, we observed that the expression of lung tight junction proteins, which was significantly down-regulated by leflunomide treatment, was effectively restored in USP8^fl/fl; Sftpc-CreERT2 mice (Supplementary Fig. 3B). Interestingly, however, USP8 deficiency failed to attenuate cell death or rescue cell viability in AT2 cells subjected to RSL3-induced stress (Supplementary Fig. 3C and D). Concurrently, USP8 knockout was unable to suppress the upregulation of ferroptosis-related proteins triggered by RSL3 stimulation (Supplementary Fig. 3E). Consistent with these in vitro findings, the inability of USP8 depletion to protect against RSL3-induced barrier disruption was successfully recapitulated in our in vivo animal model (Supplementary Fig. 3F). These observations prompted us to examine whether the anti-ferroptotic effect of USP8 knockdown in lung epithelial cells is associated with mitochondrial regulation. However, live-cell mitochondrial staining of AT2 cells revealed that USP8 depletion did not significantly affect mitochondrial mean branch length or mean aspect ratio (Fig. 2F and Supplementary Fig. 3G). We therefore examined the expression of genes involved in mitochondrial biogenesis and dynamics, including PGC-1α as well as the fission- and fusion-related genes Drp1, OPA1, Mfn1, and Mfn2. Notably, USP8 knockdown selectively restored PGC-1α expression, while having little effect on the expression of other mitochondrial dynamics–related genesh. This is highly consistent with our previous findings that USP8 altered neither mitochondrial branch length nor cross-sectional area (Supplementary Fig. 3H). Our previous studies have reported that PGC-1α expression in macrophages initially increases and subsequently decreases following LPS stimulation [11]. Consistently, PGC-1α levels were markedly reduced during the late phase of LPS stimulation in AT2 cells, whereas knockout of USP8 effectively rescued PGC-1α expression (Fig. 2G) [12]. Given that PGC-1α is a key transcriptional coactivator that promotes mitochondrial biogenesis [13], we labeled functional mitochondria with mitoGFP and transferred them into AT2 cells derived from USP8-deficient mice (Fig. 2H–I). Importantly, cytosolic mitochondrial DNA (mtDNA) levels remained unchanged before and after mitochondrial transfer (Fig. 2J). Finally, analysis of ferroptosis-related genes at both the protein and mRNA levels demonstrated that exogenous mitochondrial supplementation synergized with USP8 knockdown to suppress ferroptosis (Fig. 2K–L and Supplementary Fig. 3I–J).

Fig. 2.

Knockdown of USP8 elevates PGC-1α expression and attenuates ferroptosis in AT2 cells. (A) Schematic illustration of the molecular targets of ferroptosis inducers. (B) Effects of USP8 depletion on Leflunomide-induced ferroptosis in AT2 cells. (C) Changes in intracellular GSH and MDA levels in USP8^fl/+ and USP8^fl/fl; Sftpc-Cre mice following Leflunomide treatment. (n = 6 mice per group) (D) Assessment of cell viability under different treatment conditions. (n = 6 independent experiments) (E) Quantification of cell death (%) under different treatment conditions. (n = 6 independent experiments) (F) MitoTracker staining showing mitochondrial morphology in AT2 cells. (Scale bar = 10 μm) (G) Western blot analysis of PGC-1α protein levels in wild-type and USP8^fl/fl; Sftpc-Cre mice. (H) Schematic illustration of the mitochondrial transfer strategy. (I) MitoGFP labeling to evaluate mitochondrial transfer efficiency. (Scale bar = 10 μm) (J) Measurement of cytosolic mitochondrial DNA (mtDNA) levels. (n = 6 independent experiments) (K) Western blot analysis of ferroptosis-related proteins under different treatment conditions. (L) Immunohistochemical (IHC) analysis evaluating 4-HNE expression levels across different treatment groups following lipopolysaccharide (LPS) stimulation. (n = 6 independent experiments).

3.3. USP8 directly interacts with the PGC-1α domain to exert regulatory effects

Our previous results demonstrated that USP8 regulates the expression of PGC-1α. To determine whether this regulation involves a direct interaction between USP8 and PGC-1α, we performed a series of binding and localization analyses. We first examined the interaction between USP8 and PGC-1α across different lung-related cell types and found that this association was most prominent in AT2 cells and A549 cells; therefore, these two cell models were selected for subsequent experiments (Fig. 3A). Classically, USP8 is primarily localized in the cytoplasm, with a minor fraction associated with membrane compartments. Intriguingly, our cellular fractionation assay revealed that LPS challenge induced a marked nuclear translocation of USP8 (Supplementary Fig. 4A). co-immunoprecipitation analysis demonstrated that the physical interaction between USP8 and PGC-1α was substantially enhanced under LPS-stimulated conditions in AT2 and A549 cells (Fig. 3B). Consistently, Immunofluorescence co-staining further confirmed that USP8 prominently co-localized with PGC-1α in the nucleus upon LPS stimulation (Fig. 3D and Supplementary Fig. 4B). These results suggest that LPS-induced nuclear accumulation of USP8 promotes its interaction with PGC-1α. To map the region of PGC-1α responsible for USP8 binding, a series of PGC-1α truncation constructs were generated. We found that USP8 predominantly interacted with the 565–798 amino acid region of PGC-1α, which has been reported to be enriched in RS amino acid sequences (Fig. 3C and E). To further determine whether this interaction depends on the RS-rich region, we generated deletion and mutation constructs targeting this sequence. Deletion or mutation of the RS-rich region abolished the interaction between PGC-1α and USP8 (Fig. 3F and G). To establish a causal and dose-dependent relationship between USP8 and mitochondrial homeostasis, we utilized a Dox-inducible shRNA system to precisely titrate the expression of USP8. By modulating Dox concentrations, we successfully achieved a graded knockdown of USP8 at approximately 50%, 70%, and 90% at the protein level, providing a robust model to investigate its dose-sensitive effects on cellular senescence (Fig. 3H). Moreover, both overexpression and inhibition of USP8 revealed that USP8 abundance markedly affected PGC-1α protein levels without altering its mRNA expression, indicating that USP8 regulates PGC-1α primarily at the post-transcriptional level (Fig. 3I).

Fig. 3.

USP8 directly interacts with the 565–798 RS-rich region of PGC-1α (A) Interaction between USP8 and PGC-1α was examined in different cell lines by co-immunoprecipitation analysis. (B) The interaction between USP8 and PGC-1α following LPS stimulation was assessed by co-immunoprecipitation. (C) Schematic representation of the PGC-1α protein structure and truncation constructs. (D) Immunofluorescence staining showing the colocalization of USP8 and PGC-1α in AT2 cells after LPS stimulation. (Scale bar = 10 μm) (E) Different PGC-1α truncation plasmids were constructed and transfected into HEK293T cells to map the specific region responsible for USP8 binding. (F) Schematic alignment of amino acid sequences of PGC-1α from different species, highlighting conserved RS-rich regions. (G) The interaction between USP8 and PGC-1α was evaluated after deletion or mutation of the RS-rich amino acid sequence. (H) USP8 was knocked down in AT2 cells at graded efficiencies (approximately 50%, 70%, and 90%) by modulating Dox concentrations, followed by analysis of PGC-1α protein and mRNA levels (I) USP8 was overexpressed in AT2 cells, and the protein and mRNA levels of PGC-1α were subsequently determined. (n = 6 independent experiments).

3.4. USP8 regulates PGC-1α in a deubiquitinase-independent manner

To further elucidate the mechanism by which USP8 regulates PGC-1α, cycloheximide (CHX), a widely used protein synthesis inhibitor, was applied to assess PGC-1α protein stability. CHX chase assays revealed that simultaneous knockdown of USP8 significantly delayed the degradation of PGC-1α, indicating that USP8 affects PGC-1α protein turnover [14] (Fig. 4A and B). In parallel, treatment with the proteasome inhibitor MG132 effectively blocked PGC-1α degradation, suggesting that PGC-1α is primarily degraded via the proteasomal pathway (Fig. 4C). Moreover, we observed a progressive increase in PGC-1α ubiquitination following prolonged LPS stimulation (Fig. 4D). Given that USP8 is a well-characterized deubiquitinase reported to regulate ubiquitin-dependent processes in multiple disease models, we next examined whether its effect on PGC-1α depends on its catalytic activity [15]. Unexpectedly, deletion of USP8 resulted in a marked reduction in PGC-1α ubiquitination levels, prompting us to investigate a potential deubiquitinase-independent mechanism (Fig. 4E). To this end, a catalytically inactive USP8 mutant (USP8^C786G) was generated as previously described [16]. Notably, both wild-type USP8 and USP8^C786G exerted comparable effects on PGC-1α ubiquitination (Fig. 4F and G). Furthermore, co-immunoprecipitation assays demonstrated that USP8^C786G retained the ability to interact with PGC-1α at both endogenous and exogenous levels, independent of its deubiquitinase activity (Fig. 4H and I). To extend these findings in vivo, USP8^C786G was reintroduced into USP8^fl/fl; Sftpc-Cre mice via AAV-mediated gene delivery (Fig. 4J). Re-expression of USP8^C786G largely abolished the protective phenotype conferred by USP8 deletion, as evidenced by aggravated lung injury and restored pathological features (Fig. 4K–M). Consistently, the LPS-induced expression of ferroptosis-related genes at both the protein and mRNA levels was also restored, thereby reversing the inhibitory effect of USP8 deletion on ferroptosis (Fig. 4N and O). Collectively, these results demonstrate that USP8 interacts with and regulates PGC-1α in a manner that is independent of its deubiquitinase activity.

Fig. 4.

USP8 regulates PGC-1α independently of its deubiquitinase activity (A–B) USP8 was knocked down in AT2 cells, followed by treatment with cycloheximide (CHX, 200 μg/mL) for the indicated times to assess PGC-1α protein stability. PGC-1α band intensities were quantified using ImageJ, normalized to GAPDH, and further normalized to the t = 0 time point. (C) AT2 cells were treated with MG132 (10 μM), NH₄Cl (100 nM), or 3-MA (5 μM) for 12 h, followed by quantification of PGC-1α expression. (D) Western blot analysis of PGC-1α ubiquitination levels following LPS stimulation for the indicated durations. (E) PGC-1α ubiquitination levels were examined after USP8 deletion in AT2 and A549 cells. (F) Schematic illustration of the construction of the catalytically inactive USP8 mutant (USP8^C786G). (G) The effects of USP8^C786G on PGC-1α ubiquitination were evaluated in AT2 and A549 cells. (H) Coomassie Blue staining was used to assess the in vitro binding between USP8^C786G and PGC-1α. (I) The interaction between USP8^C786G and PGC-1α was examined in HEK293T cells. (J) Schematic diagram illustrating the in vivo experimental design for USP8^C786G re-expression. (K) Representative hematoxylin and eosin (H&E) staining images showing alveolar structures in mice subjected to different treatments. (Scale bar = 50 μm) (L) Quantitative assessment of lung injury scores. (n = 6 mice per group) (M) Survival curves of USP8^fl/fl; Sftpc-Cre and USP8^C786G mice following LPS challenge. (n = 20 mice per group) (N–O) Protein (N) and mRNA (O) levels of ferroptosis-related genes in lung tissues under different experimental conditions. (n = 6 mice per group).

3.5. USP8 recruits USP24 to regulate PGC-1α deubiquitination

To determine whether USP8 exerts its regulatory function by recruiting other deubiquitinases, USP8 was depleted in AT2 cells and subsequently reconstituted with either wild-type USP8 (WT) or a catalytically inactive mutant. Consistent with the in vivo observations, re-expression of either USP8-WT or the mutant form similarly increased the protein and mRNA levels of ferroptosis-related genes (Fig. 5A and B), suggesting that USP8 may regulate this process through interaction with additional deubiquitinases rather than relying on its intrinsic enzymatic activity. To identify potential deubiquitinases associated with USP8, IP-MS analysis was performed (Fig. 5C). Among the candidates, USP24 showed a robust interaction with USP8, which was further validated by co-immunoprecipitation in both A549 and AT2 cells, whereas other deubiquitinases displayed weak or no detectable binding (Supplementary Fig. 4C). Notably, deletion of USP8 markedly enhanced the interaction between USP24 and PGC-1α, as demonstrated by co-immunoprecipitation and immunofluorescence assays (Fig. 5D–F and Supplementary Fig. 4D). Concurrently, our in vitro investigations demonstrated that USP24 robustly interacts with PGC-1α, as evidenced by in vitro binding assays (Supplementary Fig. 4E). Consistently, increased USP24–PGC-1α interaction was also observed in lung tissues from USP8^fl/fl; Sftpc-CreERT2 mice (Fig. 5G). In vitro GST pull-down assays further confirmed that increasing amounts of USP8 progressively reduced the binding of USP24 to PGC-1α (Fig. 5H). Moreover, reconstitution of USP8-deficient AT2 and A549 cells with either USP8-WT or USP8^C786G similarly disrupted the interaction between USP24 and PGC-1α, indicating a competitive binding mechanism independent of USP8 catalytic activity (Fig. 5I and Supplementary Fig. 4F). These findings were further supported by exogenous expression assays in HEK293T cells (Fig. 5J). In contrast, deletion of USP24 did not affect the interaction between USP8 and PGC-1α (Fig. 5K).

Fig. 5.

USP8 competitively interacts with USP24 and PGC-1α. (A–B) AT2 cells with USP8 depletion were reconstituted with wild-type USP8 or the catalytically inactive mutant USP8^C786G, followed by analysis of ferroptosis-related protein (A) and mRNA (B) expression. (n = 6 independent experiments) (C) Schematic workflow of the IP-MS strategy used to identify USP8-interacting deubiquitinases. (D) Interaction between PGC-1α and USP24 in USP8-deficient AT2 cells. (E) Interaction between PGC-1α and USP24 in USP8-deficient A549 cells. (F) Immunofluorescence analysis showing the colocalization of PGC-1α and USP24. (Scale bar = 10 μm) (G) Association of PGC-1α with USP24 in lung tissues from USP8^fl/fl and USP8^fl/fl; Sftpc-Cre mice. (H) In vitro GST pull-down assays examining the interaction between PGC-1α and USP24 under different USP8 conditions. (I) Interaction between PGC-1α and the catalytically inactive USP8^C786G mutant in USP8-deficient AT2 cells. (J) Exogenous validation of the interaction between PGC-1α and USP8 (or USP8^C786G) in HEK293T cells. (K) Interaction between USP8 and PGC-1α following USP24 depletion.

3.6. USP24 deubiquitinates PGC-1α by removing K27-linked ubiquitin chains at K777, K779, and K790

To further assess the functional role of USP24 in regulating PGC-1α, USP24-deficient mice were subjected to LPS challenge. Loss of USP24 markedly enhanced the inhibitory effect of LPS on PGC-1α expression at both the protein and mRNA levels (Fig. 6A and B). Consistently, overexpression of USP24 in AT2 cells significantly reduced LPS-induced ubiquitination of PGC-1α (Fig. 6C). To determine the ubiquitin linkage specificity involved, ubiquitin mutants were introduced, revealing that only mutation of K27 (K27R) abolished the deubiquitinating effect of USP24 on PGC-1α (Fig. 6D). Mapping analyses further demonstrated that USP24 binds to the 565–798 amino acid region of PGC-1α, similar to the binding region identified for USP8, supporting a competitive interaction model (Fig. 6E and F and Supplementary Fig. 5A). To evaluate the requirement of USP24 enzymatic activity, a catalytically inactive mutant (USP24^C1695A) was generated, which failed to deubiquitinate PGC-1α (Fig. 6G and H). To identify the specific lysine residues targeted by USP24, multiple PGC-1α lysine mutants were constructed (Supplementary Fig. 5B). Simultaneous mutation of K777, K779, and K790 completely abolished the deubiquitinating effect of USP24, whereas single-site mutations partially attenuated this effect (Supplementary Fig. 5C–D). To map the functional domain, we generated a truncation mutant (PGC-1α^Δ565–798, designated PGC-1α^Del). PGC-1α^Del exhibited comparable stability to wild-type PGC-1α^Del (Supplementary Fig. 5E). Notably, expression of PGC-1α^Del blunted the anti-ferroptotic protection conferred by USP8 knockout (Supplementary Fig. 5F–G). Furthermore, PR-619 treatment reduced PGC-1α expression, whereas USP8 depletion increased PGC-1α fluorescence intensity. This accumulation was completely abolished by PGC-1α^Del. Importantly, these effects occurred strictly under LPS challenge rather than basal conditions (Fig. 6I and Supplementary Fig. 5H).

Fig. 6.

USP24 regulates PGC-1α by removing K27-linked ubiquitin chains. (A–B) Immunohistochemical staining of PGC-1α in lung tissues, with quantitative analysis shown in the bar graph. (n = 6 mice per group) (C) Ubiquitination levels of PGC-1α in AT2 cells overexpressing USP24 following LPS stimulation at the indicated time points. (D) Deubiquitination of PGC-1α by USP24 following transfection with ubiquitin mutants carrying different linkage-site substitutions in HEK293T cells. (E) Schematic representation of the PGC-1α structural regions. (F) Mapping of the USP24–PGC-1α interaction using a series of PGC-1α truncation mutants in HEK293T cells. (G) Assessment of whether PGC-1α deubiquitination depends on USP24 catalytic activity using the enzymatically inactive USP24 mutant (USP24^C1695A) in HEK293T cells. (H) Immunoblot analysis showing competitive interactions between USP24 and wild-type USP8 or the USP8^C786G mutant in HEK293T cells. (I) Immunofluorescence analysis of PGC-1α fluorescence intensity under the indicated experimental conditions. (Scale bar = 5 μm).

3.7. Loss of USP24 abolishes the protective effect of USP8 deficiency against lung injury

To further investigate the role of USP24 in vivo, we generated global USP24 knockout mice and confirmed the deletion efficiency in lung tissues (Fig. 7A and Supplementary Fig. 6A–B), and subsequently crossed USP24^−/− mice with USP8^fl/fl; Sftpc-CreERT2 to obtain USP8^fl/fl; Sftpc-CreERT2; USP24^−/− (USP8^fl/flSft^CreUSP24^−/−) mice. In the in vivo ALI model, loss of USP24 markedly abolished the protective effects conferred by USP8 deficiency (Fig. 7B and Supplementary Fig. 6C–D). Consistently, dual deletion of USP24 and USP8 in AT2 cells (sgDual) eliminated the beneficial effects observed after USP8 loss alone: while USP8 deletion restored LPS-induced downregulation of lung barrier–associated genes, suppressed the upregulation of ferroptosis-related genes, and increased PGC-1α expression, these protective effects were no longer evident in the absence of USP24 (Fig. 7C). Immunoblot analysis of ferroptosis-related proteins in mouse lung tissues further supported these findings, revealing that LPS challenge markedly induced the upregulation of USP24 (Fig. 7G). Collectively, these results indicate that USP8 depletion rescues LPS-induced lung epithelial injury in a USP24-dependent manner. Moreover, AAV-mediated overexpression of USP24 in USP8^fl/fl; Sftpc-CreERT2 mice (Fig. 7D) synergized with USP8 deficiency to protect against acute lung injury, as evidenced by improved respiratory function (Fig. 7E–F and Supplementary Fig. 6E–F). Concurrently, genetic depletion of USP24 effectively abrogated the protective effects on the lung epithelium originally afforded by USP8 deficiency, primarily by suppressing the USP8 depletion–induced upregulation of PGC-1α (Supplementary Fig. 6G–H). Together, these results indicate that USP8 competitively interferes with the interaction between USP24 and PGC-1α, thereby reducing PGC-1α levels, activating ferroptosis, and disrupting the pulmonary epithelial barrier, whereas upon USP8 deletion, USP24 directly interacts with and deubiquitinates PGC-1α, leading to increased PGC-1α levels, suppression of ferroptosis, and restoration of epithelial barrier integrity (Fig. 7H).

Fig. 7.

In vivo validation that the protective effect of USP8 depletion depends on USP24. (A) Schematic illustration of the generation of whole-body USP24^−/− mice. (B) Representative H&E staining of lung tissues from Sftpc-CreERT2 (Sft^Cre), USP8^fl/fl; Sftpc-CreERT2 (USP8^fl/flSft^Cre), USP24^−/−, USP8^fl/fl; Sftpc-CreERT2; USP24^−/− (USP8^fl/flSft^CreUSP24^−/−), and the indicated compound mutant mice. (C) RT–qPCR analysis of mRNA levels of lung epithelial barrier–related genes, ferroptosis-associated genes, and mitochondrial-related genes in AT2 cells under the indicated conditions. (n = 3 independent experiments). (D) Schematic illustration of AAV-mediated USP24 overexpression in mice. (E) Representative H&E staining showing lung histopathological changes in mice subjected to the indicated treatments. (F) Assessment of respiratory function in mice under the indicated experimental conditions. The Sftpc-CreERT2 (Sft^Cre) group received saline, whereas all other groups were challenged with LPS. (n = 6 mice per group) (G) Immunoblot analysis of ferroptosis-related protein expression in mouse lung tissues. (H) Schematic model illustrating the proposed mechanism by which USP8 regulates PGC-1α and ferroptosis through USP24 during acute lung injury.

4. Discussion

USP8 is a well-characterized deubiquitinating enzyme that has been extensively reported in multiple disease models [17,18]. In this study, we demonstrate that deletion of USP8 in the lung restores LPS-induced disruption of the pulmonary epithelial barrier and suppresses ferroptosis. Mechanistically, this protective effect is associated with regulation of the mitochondrial biogenesis regulator PGC-1α. Notably, expression of a catalytically inactive US8^C786G (USP8 mutant) phenocopied the effects of wild-type USP8, suggesting that USP8 functions in acute lung injury in a manner independent of its deubiquitinating activity, likely as a scaffolding or recruitment factor. Using IP–MS and co-immunoprecipitation analyses, we further identified USP24 as a shared interactor of USP8 and PGC-1α. Both in vivo and in vitro experiments establish USP24 as a critical mediator of USP8-dependent regulation of PGC-1α, as loss of USP24 exacerbated acute lung injury and abolished the protective effects conferred by USP8 deficiency. Moreover, we show that USP8 competes with USP24 for binding to PGC-1α, and that USP8 and USP24 can also directly interact with each other.

Acute lung injury is widely attributed to LPS-induced activation of TLR4 signaling, macrophage activation, and consequent disruption of epithelial barrier integrity [19,20]. Our previous work demonstrated that early intervention targeting mitochondrial homeostasis in macrophages can attenuate pulmonary inflammation, largely through modulation of inflammatory cytokine release [6]. PGC-1α is a central regulator of mitochondrial biogenesis and has been implicated in multiple LPS-induced organ injury models, including liver injury, where modulation of PGC-1α or the Sirt1/PGC-1α signaling axis confers tissue protection [21,22]. Dysregulated mitochondrial biogenesis is closely linked to activation of ferroptosis pathways, a process that has been extensively characterized [23]. Consistent with these findings, we show that USP8-mediated regulation of PGC-1α attenuates ferroptosis and mitigates LPS-induced acute lung injury. PGC-1α contains three structural domains, and our data indicate that USP8 interacts with an RS amino acid–rich region rather than other domains. Intriguingly, whereas USP8 is conventionally recognized as a predominantly cytoplasmic enzyme under basal conditions, we observed that LPS challenge triggers its prominent nuclear translocation, facilitating its physical interaction with nuclear PGC-1α. Distinct from its classical deubiquitinating function, however, our results demonstrate that USP8 acts independently of its catalytic activity in this context [24]. Instead, nuclear USP8 functions as a molecular scaffold that recruits USP24, which in turn mediates the deubiquitination and stabilization of PGC-1α to protect against ALI. A similar non-canonical role has been reported for USP25 in non–small cell lung cancer, where it lacks deubiquitinating activity and instead recruits RNF31 to influence tumorigenesis [25]. USP24 is a broadly expressed deubiquitinating enzyme with context-dependent roles in ferroptosis across disease models. For example, USP24 has been reported to exacerbate diabetic cardiomyopathy via activation of NF-κB signaling [26], whereas more recent studies demonstrate that USP24 suppresses ferroptosis and inhibits triple-negative breast cancer progression by stabilizing DHODH [27]. These findings underscore the functional plasticity of USP24. In the present study, we identify USP24 as a protective factor in acute lung injury, acting through deubiquitination and stabilization of PGC-1α to suppress ferroptosis. Intriguingly, USP24 also binds to the RS amino acid–rich region of PGC-1α, and directly interacts with USP8, supporting a model in which USP8 competitively regulates the USP24–PGC-1α interaction to influence disease progression.

Together, our findings reveal a previously unrecognized mechanism in LPS-induced acute lung injury whereby USP8 functions independently of its deubiquitinating activity to modulate mitochondrial homeostasis and ferroptosis through USP24 and PGC-1α, providing new mechanistic insight into the pathogenesis of acute lung injury. Nevertheless, this study has limitations. Owing to the limited therapeutic options for sepsis-associated acute lung injury, which primarily rely on positive-pressure ventilation, access to clinical samples was not feasible, precluding validation of these mechanisms in human tissues. Future clinical studies will be required to substantiate the translational relevance of our findings. In addition, whether other deubiquitinating enzymes play analogous non-canonical roles in acute lung injury remains to be determined.

In summary, we demonstrate that USP8 regulates acute lung injury through a deubiquitination-independent mechanism by coordinating USP24- and PGC-1α–dependent control of ferroptosis. Upon USP8 loss, USP24 directly associates with and deubiquitinates PGC-1α, elevating its protein levels and suppressing downstream ferroptotic signaling, thereby preserving pulmonary epithelial barrier integrity.

Ethics declaration

This study was conducted in accordance with the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines. This study was approved by the Animal Care and Use Committee of Renmin Hospital of Wuhan University. (Approval No. WDRM20220604)

Availability of data and materials

The RNA sequencing datasets generated in this study have been submitted to the Genome Sequence Archive (GSA) under accession number CRA047018 and are accessible to the public upon publication. The mass spectrometry proteomics datasets have been uploaded to the ProteomeXchange Consortium through the iProX repository (dataset ID: PXD081572). Additional experimental raw data supporting the findings of this study are available from the corresponding author upon reasonable request.

Consent for publication

Not Applicable.

Clinical trial number

Not Applicable.

Funding

This work was supported by the National Natural Science Foundation of China, China (210971479, 220172497) and Natural Science Foundation of Hubei, China (2025AFC074).

CRediT authorship contribution statement

Wei Lai: Conceptualization, Data curation, Formal analysis. Xin Xing: Formal analysis, Investigation, Methodology, Project administration. Songhela Ahan: Investigation, Methodology, Resources. Yan Liu: Methodology, Project administration, Resources. Wanli Jiang: Funding acquisition, Project administration, Software, Supervision.

Declaration of competing interest

The authors declared that we have no financial or non-financial conflicts of interest regarding the content of this manuscript.

Appendix A. Supplementary data

The following are the Supplementary data to this article: