Nedd4L-mediated stabilization of morphogenetic protein receptor 2 in lung fibroblasts confers protection against pulmonary fibrosis

Abstract

Idiopathic pulmonary fibrosis (IPF) is an interstitial lung disease that advances over time and is ultimately life-threatening. Despite advances in antifibrotic therapies, disease progression remains largely unavoidable, highlighting the necessity for advancing mechanistic research and identifying new therapeutic options. Bone morphogenetic protein receptor 2 (BMPR2) is a key receptor for BMP signaling and exerts potent antifibrotic role in lung fibrosis; however, the molecular mechanisms regulating BMPR2 protein stability in fibrotic lungs remain poorly understood. In this study, we identify that the E3 ubiquitin ligase neural precursor cell expressed developmentally downregulated protein 4-like (Nedd4L) as a critical regulator of BMPR2 stability in lung fibroblasts. BMPR2 levels are markedly reduced in lung fibroblasts derived from patients with IPF and in lung tissues from experimental lung fibrosis models. Mechanically, elevated lipid peroxidation in fibrotic lungs induces K48-linked ubiquitination of BMPR2, leading to its degradation and attenuation of BMPs-induced Smad1/5 phosphorylation. Notably, Nedd4L overexpression counteracts lipid peroxidation-induced BMPR2 degradation and promotes BMPs/BMPR2 signaling by shifting BMPR2 ubiquitination toward a stabilizing K63-linked polyubiquitination in lung fibroblasts. We further show that Nedd4L is regulated through threonine phosphorylated by Raf1 kinase in response to increased lipid peroxidation. In vivo, fibroblast-Nedd4L-deficient mice exacerbates experimental lung fibrosis, which is significantly alleviated by BMPR2 overexpression. Collectively, our findings reveal a previously unrecoganized regulatory mechanism of BMPR2 degradation in fibrotic lungs and indicate that stabilizing BMPR2 in lung fibroblasts may represent a promising approach for the treatment of lung fibrosis.

Introduction

Idiopathic pulmonary fibrosis (IPF) is a life-threatening interstitial lung disease driven by excessive extracellular matrix (ECM) deposition within the lungs,^1,2 leading to tissue stiffening and thickening, ultimately impairing gas exchange.³ Following diagnosis, the median life expectancy for patients with IPF is estimated to be between 3 and 5 years.^1,4 Despite FDA-approved therapies that limit the rate of progression, no current treatments can cure the disease or restore lung function.⁵ Understanding the molecular mechanisms underlying IPF pathogenesis is essential for advancing novel therapeutic strategies for IPF and other fibrotic lung diseases.

Activated fibrotic fibroblasts are key drivers of IPF pathogenesis by producing ECM proteins that accumulate in lung interstitial tissues and disrupt alveolar architecture and lung function.^6–8 Benefiting from the rapid development of single-cell RNA sequencing, additional subtypes of lung fibroblasts, in addition to alveolar fibroblasts and myofibroblasts, have been identified. These include lipofibroblasts and Ebf1+ fibroblasts, which are implicated in the onset and advancement of pulmonary fibrosis. Thus, inhibiting abnormal fibroblast activation is a promising approach for treating IPF. TGF-β1 signaling is a key driver of α-smooth muscle actin (α-SMA), collagen, elastin, and fibronectin (FN) upregulation.⁹ Bone morphogenetic proteins (BMPs), particularly BMP4 and BMP7, have been shown to counteract fibrotic signaling.^10–15 BMP signaling induces Smad1/5/8 phosphorylation, thereby counteracting the TGF-β/Smad2/3 pathway through BMP receptor activation.^16,17 BMPR2, a type II BMP receptor, is broadly expressed throughout lung cell populations, with particularly high levels in pulmonary endothelial cells and fibroblasts. BMPR2 has been widely investigated in pulmonary artery hypertension (PAH).^18–20 Dysfunction of BMPR2 due to its mutation is a major driver of PAH pathogenesis.^18–20 Recent emerging evidence suggests that BMPR2 also contributes to the development of IPF.^21–23 Notably, BMPR2 is markedly reduced in lungs from IPF patients and in bleomycin-induced experimental lung fibrosis.^16,21,24 However, the regulatory mechanisms controlling BMPR2 expression during fibrosis remain poorly understood. Recent studies highlight lipid peroxidation as a critical driver of IPF progression. Lipid peroxidation is initiated when free radicals attack polyunsaturated fatty acids,²⁵ contributing to TGF-β1 signaling and ECM upregulation.^26–28 Increased lipid peroxidation has been observed in both IPF patient lungs and bleomycin-induced mouse fibrotic lungs,^29,30 reflecting a disturbance in the equilibrium between oxidative production and the antioxidant protection. The molecular interplay between lipid peroxidation and the BMPR2 pathway has not been revealed.

To investigate the molecular regulation of BMPR2 in fibrotic fibroblasts, we focused on determining its stability through the ubiquitination-mediated protein degradation system, which is essential for preserving intracellular protein homeostasis.^31–35 Our previous work has shown that the E3 ubiquitin ligase Nedd4L in lungs was significantly decreased in IPF patients and bleomycin-challenged mice, and that decreased Nedd4L expression was associated with increased disease severity. Overexpression of Nedd4L suppressed TGF-β1-induced myofibroblast transition from fibroblasts and mitigated fibrotic changes in mice exposed to bleomycin.³¹ Additionally, Duerr et al. found that Nedd4L suppression in lung epithelial cells contributes to pulmonary fibrosis,³² suggesting a broader role of Nedd4L in IPF pathogenesis. However, the role of Nedd4L in the BMP/BMPR2 signaling axis remains unclear.

Here, we aimed to characterize the molecular mechanisms that regulate BMPR2 stability in lung fibroblasts, focusing on the interplay between lipid peroxidation and Nedd4L-mediated BMPR2 stabilization. We demonstrate that lipid peroxidation reduces BMPR2 stability in lung fibroblasts, where Nedd4L promotes BMPR2 protein stability and function via K63-linked polyubiquitination. Furthermore, administration of BMPR2 to fibroblast-specific Nedd4L-deficient mice alleviated bleomycin-induced fibrosis. These findings provide new insights into BMPR2 regulation and suggest that stabilizing BMPR2 could represent a promising therapeutic approach for IPF.

Results

BMPR2 protein expression is markedly diminished in fibrotic lungs

Previous findings suggest that BMPR2 protein expression is reduced in IPF lung tissues^16,21,24; however, the molecular mechanisms that regulate BMPR2 in lung fibroblasts remain undefined. To explore the role of BMPR2 in lung fibrosis, we first assessed BMPR2 expression in primary lung fibroblasts isolated from individuals with IPF and compared with those from healthy donors. BMPR2 protein abundance was markedly reduced in IPF lung fibroblasts (IPF-LF) relative to lung fibroblasts from control subjects (HLF) (Fig. 1a), whereas the mRNA expression of BMPR2 remained unchanged (Fig. 1b), indicating that BMPR2 downregulation in lung fibroblasts from IPF patients is likely driven by post-transcriptional mechanisms, such as increased protein degradation. This conclusion was further supported by our reanalysis of publicly available datasets (https://www.ncbi.nim.nih.gov/geo/) comparing BMPR2 transcript levels in lung tissues and primary lung fibroblasts from healthy individuals and IPF patients, and the data showed that no gene levels were significantly different between these two groups (Fig. 1c, d). Furthermore, BMPR2 protein levels in lungs from bleomycin-challenged mice were analyzed by immunoblotting. BMPR2 protein levels in whole lung tissues were markedly reduced after 1, 2, and 3 weeks of bleomycin administration compared to those in PBS control mice (Fig. 1e, h, and Supplementary Fig. 1a). Flow cytometry examining BMPR2 levels in lung fibroblasts isolated from mice treated with PBS and bleomycin demonstrated that BMPR2 protein levels were downregulated as well (Supplementary Fig. 1c). BMPR2 mRNA expression, however, was not significantly changed (Fig. 1f, and Supplementary Fig. 1b), and these findings were confirmed by reanalysis of the available dataset (Fig. 1g). Furthermore, we employed a repetitive bleomycin model to validate the data obtained from the acute fibrosis model and demonstrated a reduction in BMPR2 expression in fibrotic lungs in a chronic experimental fibrosis model (Fig. 1i). After 1 week of intratracheal LPS administration, postinjury fibrotic responses are induced in the lungs during the repair and remodeling phase.^36,37 BMPR2 levels were significantly reduced in the fibrotic phase following lung injury (Fig. 1j). Collectively, these results demonstrate that the reduction in BMPR2 in fibrotic fibroblasts does not stem from transcriptional alterations but is instead closely linked to impaired BMPR2 protein stability.

Fig. 1.

BMPR2 levels are decreased in lung fibroblasts from IPF lungs and mouse lungs of experimental lung fibrosis. a BMPR2 protein levels in normal human lung fibroblasts (HLFs) (n = 9) and IPF lung fibroblasts (IPF-LF) (n = 11) were analyzed by immunoblotting. Representative samples are shown (n = 3–5). The intensities of BMPR2 were analyzed by ImageJ. b mRNA expression of BMPR2 was detected by real-time PCR in HLF (n = 3) and IPF-LF (n = 5). Dots shown are duplications of each sample. c Reanalysis of the GDS4279 dataset to examine BMPR2 gene expression in the lung tissues of normal subjects and IPF patients. d Reanalysis of the GDS4580 dataset to compare BMPR2 gene expression in lung fibroblasts from normal subjects and IPF patients. e BMPR2 protein levels in PBS- or bleomycin (2.2 U/kg, 2 weeks)-challenged C57BL/6J wild-type mouse lung tissues were analyzed by immunoblotting. The intensities of BMPR2 (n = 4–6) were quantified by ImageJ. f mRNA expression of BMPR2 from experimental PF mouse lungs was detected by real-time PCR (n = 4–6). g Reanalysis of GDS4902 and GDS5078 datasets to compare BMPR2 gene levels in PBS- or bleomycin-challenged mice. h BMPR2 protein levels in lung tissues from PBS- or bleomycin (2.2 U/kg, 3 weeks)-challenged C57BL/6J wild type mice were analyzed by immunoblotting. The intensities of BMPR2 (n = 6) were quantified by ImageJ. i BMPR2 protein levels in lung tissues from PBS- or repetitive bleomycin (1.2 U/kg, 6-instillation, two-week intervals)-challenged C57BL/6J wild-type mice were analyzed by immunoblotting. j BMPR2 protein levels in lung tissues from PBS- or LPS (i.t. 5 mg/kg, 7 days)-challenged C57BL/6J wild type mice were analyzed by immunoblotting. The intensities of BMPR2 (n = 3–4) were quantified by ImageJ. ^**p < 0.01; ns, not significant

Elevated lipid peroxidation in PF mediates BMPR2 degradation

Lipid peroxidation is a series of reactions of oxidative degradation of lipids that results in the formation of lipid peroxides and causes cell dysfunction.²⁵ It has been shown to be elevated in lung tissues from IPF patients and in bleomycin-induced PF.^29,30 Elevated lipid peroxidation promotes the upregulation of α-SMA and collagen.^26–28 Lipid peroxidation levels can be assessed by quantifying its reactive end products, including malondialdehyde (MDA) and 4-hydroxy-2-nonenal (4-HNE).³⁸ We confirmed the increased lipid peroxidation with an MDA assay and ATF4 immunostaining of mouse and human fibrotic lungs (Supplementary Fig. 2a–d). Bleomycin challenge-induced lung fibrosis was evaluated by histological staining (Supplementary Fig. 2e). ATF4, a transcription factor mechanistically associated with lipid peroxidation, was markedly upregulated in experimental PF and colocalized with α-SMA-positive fibroblasts, consistent with enhanced lipid peroxidation in fibrotic fibroblasts.

GPX4, a key antioxidant enzyme that reduces lipid peroxidation, is markedly downregulated in IPF lungs. Consistent with its protective role, GPX4-deficient mice develop more severe bleomycin-induced lung fibrosis,^26–28 suggesting that lipid peroxidation contributes to the development of lung fibrosis. We confirmed that GPX4 levels were significantly diminished in lung fibroblasts derived from both IPF and bleomycin-induced PF models (Supplementary Fig. 2d, e). We used RSL3, which inhibits GPX4 by downregulating GPX4 mRNA expression and has been experimentally used to induce lipid peroxidation, in lung fibroblasts. The effect of RSL3 on the induction of lipid peroxidation was validated in human lung fibroblasts (hLF/hTERT and IMR90) (Supplementary Fig. 3). IMR90 cells were exposed to RSL3 across a range of concentrations and time periods, and treatment with RSL3 (0.5 µM) for 3 h resulted in a markedly reduction in BMPR2 protein abundance (Fig. 2a, b). The RSL3-induced reduction in BMPR2 was confirmed in hLF/hTERT and HLF cells, as well as in myofibroblasts generated by TGF-β1-treated hLF/hTERT cells (Fig. 2c, d, and Supplementary Fig. 4a). In addition to RSL3, erastin and H₂O₂, which are also inducers of lipid peroxidation and oxidative stress, reduced BMPR2 levels in hLF/hTERT cells (Supplementary Fig. 5). TGF-β1 is well established to promote oxidative stress, and we observed that a higher concentration of TGF-β1 (20 ng/ml) reduced BMPR2 levels (Supplementary Fig. 4b). These findings suggest that increased lipid peroxide formation and oxidative stress during the progression of lung fibrosis contribute to BMPR2 reduction in lung fibroblasts. BMP4 functions as a ligand of BMPR2. RSL3 pretreatment attenuated BMP4-induced phosphorylation of Smad1/5 in IMR90 and IPF-LF cells (Fig. 2e, f), suggesting that RSL3 mitigates the BMP pathway through a reduction in BMPR2 levels.

Fig. 2.

The lipid peroxidation inducer RSL3 induces BMPR2 ubiquitination and degradation in lung fibroblasts. Immunoblotting analysis of BMPR2 in RSL3 (3 h, 0–1 μM)-treated IMR90 cells (a) and RSL3 (0.5 μM, 0-6 h)-treated IMR90 cells (b). BMPR2 intensities (n = 3) were analyzed by ImageJ. Immunoblotting analysis of BMPR2 in RSL3 (0.5 μM, 3 h)-treated hLF/hHTERT (c) and HLF cells (d). BMPR2 intensities (n = 3) were analyzed by ImageJ. IMR90 (e) or IPF-LF (f) cells were pretreated with RSL3 (0.5 μM) for 3 h, followed by BMP4 (10 ng/ml) treatment for an additional 30 min. p-Smad1/5 was analyzed by immunoblotting. The intensities of p-Smad1/5 (n = 3–4) were quantified by ImageJ. g IMR90 cells were pretreated with MG-132 (20 μM), leupeptin (100 μM), or bafilomycin A1 (BafA1, 1 nM) for 1 h, followed by RSL3 (0.5 μM) treatment for an additional 3 h. BMPR2 protein levels were detected by immunoblotting. BMPR2 intensities (n = 3–4) were analyzed by ImageJ. h K48-linked ubiquitination of BMPR2 after RSL3 (0.5 μM, 3 h) treatment was detected by a ubiquitination assay with an antibody against the K48-linked ubiquitin chain. i Fluorescence staining of DAPI (blue) and BMPR2 (green) in human lung fibroblasts (Mrc5) treated with RSL3 (0.5 μM, 1 h)/DMSO. Scale bar = 20 μm. The ratio of the intensities of the BMPR2 signal on the cell membrane and cytoplasm was analyzed by ImageJ. ^***p < 0.001, ^**p < 0.01, *p < 0.05. The immunoblots shown are representative blots from three independent experiments

Next, the molecular mechanisms underlying RSL3-induced BMPR2 degradation were investigated. Both the proteasome inhibitor MG-132 and the lysosome inhibitors leupeptin and bafilomycin A1 prevented the RSL3-induced reduction in BMPR2 (Fig. 2g), suggesting that BMPR2 degradation occurs through both the proteasome and lysosome systems. E3 ubiquitin ligases induce mono-ubiquitination, polyubiquitination, or multimono-ubiquitination of substrate proteins.³⁹ Polyubiquitin chains are formed on any of the seven lysine residues within the ubiquitin protein.⁴⁰ Among the various types of polyubiquitination, lysine (K)48- and K63-linked polyubiquitination have been the most extensively studied. K48-linked polyubiquitin chains primarily result in protein degradation.^40,41 In contrast, K63-congulated polyubiquitin chains are involved in diverse cellular processes and can, in some contexts, promote protein stabilization.^40–43 Ubiquitin assays revealed that RLS3 treatment enhanced K48-linked polyubiquitination (Fig. 2h), indicating that RSL3 induces ubiquitination-mediated BMPR2 degradation. Furthermore, fluorescence staining showed that RSL3 treatment induced BMPR2 protein internalization into the cytoplasm (Fig. 2i). These results suggest that the downregulation of BMPR2 stability and BMP/BMPR2 signaling in PF is driven by elevated lipid peroxidation during disease progression.

The E3 ubiquitin ligase Nedd4L enhances BMPR2 protein stability and facilitates activation of the BMP signaling pathway

Our previous study indicated that Nedd4L levels were diminished in lung fibroblasts derived from IPF patients and bleomycin-induced PF.³¹ To determine whether NEDD4L expression in fibrotic lung tissue is correlated with the expression of fibrotic markers, we analyzed NEDD4L, ACTA2, and COL1A2 levels in 134 IPF lung tissue samples from the Lung Genomics Research Consortium (LGRC) dataset (http://www.lung-genomics.org). The expression of NEDDL4 was negatively correlated with that of ACTA2 and COL1A2, as shown in Fig. 3a. Both ACTA2 and COL1A2 are well-known markers in fibrotic lungs. These findings further support that NEDD4L is an antifibrotic gene for lung fibrosis.

Fig. 3.

The E3 ubiquitin ligase Nedd4L stabilizes BMPR2 and modulates BMP4/7-mediated p-Smad1/5. a Gene expression data from the Lung Tissue Research Consortium (LGRC) were analyzed for NEDDL4, ACTA2 and COL1A2. The study included 134 patients with IPF. The correlations of NEDDL4 with ACTA2 and COL1A2 are shown. R² and p values from linear regression analysis are shown. b Nedd4L was knocked down with siRNA in IMR90 cells for 3 days. Cell lysates were analyzed by immunoblotting with BMPR2, Nedd4L, and β-actin antibodies. Nedd4L and BMPR2 intensities (n = 3) were analyzed by ImageJ. c Real-time PCR was performed to analyze mRNA levels in Nedd4L siRNA-transfected IMR90 cells. (n = 3). Dots shown are duplications of each sample. d siNedd4L was used to knock down the expression of Nedd4L, and cycloheximide (CHX) was utilized to examine the half-life of BMPR2 protein at 0, 2, 4, and 8 h in IMR90 cells (n = 3). BMPR2 and Nedd4L protein levels were analyzed by immunoblotting. Nedd4L was knocked down with siRNA in IMR90 (e) or IPF-LF (f) cells for 3 days, followed by BMP4 (10 ng/ml) treatment for an additional 15 min. Cell lysates were analyzed by immunoblotting. The intensities of P-Smad1/5 (n = 3) were examined by ImageJ. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001. The immunoblots shown are representative blots from three independent experiments

Given the observed reduction in BMPR2 during PF, we suspected that Nedd4L may participate in the molecular regulation of BMPR2. Knockdown of Nedd4L by siRNA transfection reduced Nedd4L levels as well as BMPR2 protein abundance (Fig. 3b), but the gene levels of BMPR2 remained unchanged (Fig. 3c). These results indicate that Nedd4L modulates BMPR2 at the protein level rather than through changes in its mRNA expression. Next, we utilized the protein synthesis inhibitor cycloheximide (CHX) to investigate whether Nedd4L regulates the stability and half-life of the BMPR2 protein. As shown in Fig. 3d, downregulation of Nedd4L significantly reduced the half-life of BMPR2, indicating that Nedd4L stabilizes BMPR2. Although Smurf1 E3 ligase has been shown to modulate BMPR2 degradation in mesenchymal stem cells,⁴⁴our data revealed that Smurf1 did not alter BMPR2 protein levels in lung fibroblasts (Supplementary Fig. 6a). Thus, we focused on the effect of Nedd4L on BMPR2 stability and BMP/BMPR2-mediated signaling in lung fibroblasts.

BMPR2 regulates BMP-induced signaling pathways.^10,16 To examine whether Nedd4L modulates BMP/BMPR2 pathway, we used BMPR2 siRNA to downregulate BMPR2 gene expression in lung fibroblasts (Supplementary Fig. 7a). BMP4-induced phosphorylation of Smad1/5 was significantly attenuated in BMPR2 siRNA-transfected lung fibroblasts (Supplementary Fig. 7b). Furthermore, we examined whether Nedd4L influences the BMP-mediated signaling pathway. BMP4- and BMP7-stimulated Smad1/5 phosphorylation was markedly reduced in Nedd4L-deficient lung fibroblasts (Fig. 3e, f, and Supplementary Fig. 7c, d). Nedd4L is expressed across all cell types (http://app.lungmap.net). However, downregulation of Nedd4L in lung epithelial cells did not alter P-Smad1/5 or BMPR2 levels (Supplementary Fig. 7e), suggesting that Nedd4L modulates BMPR2 stabilization and regulates the BMP signaling pathway in a fibroblast-specific manner.

Nedd4L targets BMPR2 for K63-linked polyubiquitination and stabilization

To investigate the mechanisms that Nedd4L promotes BMPR2 stabilization, we examined whether Nedd4L has the ability to regulate BMPR2 ubiquitination. Overexpression of Nedd4L increased K63-linked polyubiquitination of BMPR2 (Fig. 4a). Conversely, K63-linked polyubiquitination was diminished in Nedd4L-deficient cells (Fig. 4b), indicating that Nedd4L catalyzes K63-linked polyubiquitination of BMPR2.

Fig. 4.

Nedd4L stabilizes BMPR2 through K63 polyubiquitination. The Nedd4L-V5 plasmid (a) or siNedd4L (b) was transfected into IMR90 cells, followed by a ubiquitination assay with BMPR2 antibodies. Levels of K63-linked polyubiquitin chains were examined by immunoblotting. c Levels of K63-linked polyubiquitin chains were compared in BMPR2-HA wild type, BMPR2-K539R, BMPR2-K564R, and BMPR2-K673R-overexpressing HeLa cells by K63-linked ubiquitination assay. The red arrow indicates that K564R has no ability to induce K63-linked ubiquitination. d A K63-linked ubiquitination assay was performed to detect K63-linked polyubiquitination in HeLa cells cotransfected with BMPR2-HA wild type, BMPR2-K539R, or BMPR2-K564R with the Nedd4L-V5 plasmid. The red arrow indicates that K564R has no ability to induce K63-linked ubiquitination in the presence of Nedd4L-V5. e CHX was used to examine the half-life of overexpressed BMPR2-K564R and BMPR2-HA wild type in HeLa cells (n = 3). f Co-IP analysis of the association between overexpressed BMPR2-HA and Nedd4L-V5 in IMR90 cells. g The cellular localization of overexpressed Nedd4L-Xpress (green) and BMPR2-V5 (red) was detected by coimmunostaining in IMR90 cells. Scale bar = 20 μm. h Potential Nedd4L docking sites (PY motifs) on BMPR2. i Co-IP analysis of the association of Nedd4L-V5 with wild-type BMPR2-HA, BMPR2-Y432A, BMPR2-Y543AY546A, BMPR2-Y634A, and BMPR2-Y979A in HeLa cells. The red arrow indicates that Y634A-HA has a lower binding affinity for Nedd4L-V5. j Analysis of the half-life of overexpressed BMPR2-Y634A and wild-type BMPR2-HA by CHX in HeLa cells (n = 3)

According to PhosphositePlus, there are three potential ubiquitin conjugation residues (K539, K564, and K763) within BMPR2. HA-tagged BMPR2 plasmids with replacement of K to R (arginine) were generated. The K564R mutant exhibited less K63-linked polyubiquitination than the wild type and other K to R variants (Fig. 4c). Furthermore, the BMPR2-HA-K564R mutant did not display an increase in K63-linked polyubiquitination in the presence of Nedd4L-V5 (Fig. 4d), and its half-life was decreased compared to that of wild-type BMPR2 (Fig. 4e), suggesting that K564 on BMPR2 is the K63-linked ubiquitination site by Nedd4L and that K63-linked polyubiquitination facilitates the stabilization of BMPR2 protein.

Furthermore, we investigated whether Nedd4L directly mediates BMPR2 ubiquitination. A link between Nedd4L and BMPR2 was detected using a coimmunoprecipitation assay (Fig. 4f). Coimmunofluorescence staining showed that Nedd4L was colocalized with BMPR2 primarily on the plasma membrane (Fig. 4g). Moreover, we mapped the Nedd4L interaction site on BMPR2. The PY (PPXY) motif on the substrate is the Nedd4L binding domain.^45,46 BMPR2 contains five PY motifs, four of which (aa 428-432, aa 541-546, aa 632-634 and aa 977-979) are in the intracellular domain (Fig. 4h). Plasmids encoding Y → A mutations (Y432A, Y543A/Y546A, Y634A, and Y979A) were generated based on the HA-tagged BMPR2 plasmid. Co-IP showed that the binding affinity to Nedd4L was significantly reduced in the Y634A BMPR2 mutant (Fig. 4i), indicating that the PY632-634 motif on BMPR2 may be a potential binding site for Nedd4L. Furthermore, the Y634A mutant was observed to have a shorter half-life (Fig. 4j), indicating that the effect of Nedd4L on BMPR2 ubiquitination and stability occurs through direct association with BMPR2.

Nedd4L diminishes lipid peroxidation-mediated BMPR2 degradation

Next, we examined whether Nedd4L stabilizes BMPR2 in response to RLS3 treatment. Lung fibroblasts were transfected with the Nedd4L-V5 plasmid or Nedd4L siRNA, followed by treatment with RSL3. As shown in Fig. 5a, b, RSL3-induced BMPR2 degradation was rescued by Nedd4L overexpression, while it was promoted by Nedd4L knockdown. RSL3 treatment decreased K63-linked polyubiquitination, while it increased K48-conjuated polyubiquitination of BMPR2, and these effects were reversed by overexpression of Nedd4L-V5 (Fig. 5c). Taken together, these findings indicate that Nedd4L counteracts lipid peroxidation-induced BMPR2 degradation by promoting K63-linked polyubiquitination while suppressing K48-linked polyubiquitination of BMPR2, thereby enhancing its stability.

Fig. 5.

Nedd4L attenuates RSL3-induced BMPR2 reduction. a Nedd4L-V5 plasmids were transfected into IMR90 cells for 2 days, and then cells were treated with RSL3 (0.5 μM) for 3 h. Cell lysates were analyzed by immunoblotting. BMPR2 intensities (n = 3) were analyzed by ImageJ. b Nedd4L was knocked down with siRNA in IMR90 cells, and then cells were treated with RSL3 (0.5 μM) for 3 h. BMPR2 and Nedd4L protein levels were examined by immunoblotting and ImageJ (n = 3). c K63- and K48-linked polyubiquitination levels were detected by ubiquitination assay in Nedd4L-overexpressing HeLa cells treated with RSL3 (0.5 μM, 3 h). d Immunoblotting analysis of RSL3 (0.5 μM, 3 h)-treated IMR90 cells. The intensities of Nedd4L (n = 3) were analyzed by ImageJ. e Co-IP with Nedd4L antibody, followed by p-threonine immunoblotting, was performed to examine Nedd4L threonine phosphorylation by RSL3 treatment (0.5 μM) in IMR90 cells. f Nedd4L and BMPR2 were overexpressed in IMR90 cells, and then cells were stimulated with RSL3 (0.5 μM) for 0, 1, and 3 h. Co-IP analysis was performed to determine the association between Nedd4L and BMPR2. g A mass spectrometry analysis of the threonine phosphorylation of Nedd4L after RSL3 treatment (0.5 μM, 1 h) in IMR90 cells. h HLF/hTERT cells were transfected with Nedd4L-Xpress or Nedd4LT451A-Xpress with or without Raf1-V5 plasmids for 48 h. Cell lysates were subjected to IP with p-threonine antibody, followed by Xpress tag immunoblotting. ^*p < 0.05, ^***p < 0.001. The immunoblots shown are representative blots from two to three independent experiments

We further investigated the molecular mechanisms underlying the attenuation of RSL3-induced BMPR2 degradation by Nedd4L. A decreasing trend in Nedd4L was found after RSL3 treatment (Fig. 5d), while threonine, but not serine, phosphorylation of Nedd4L was observed after RSL3 treatment for 1 h (Fig. 5e, and Supplementary Fig. 8). It has been shown that phosphorylation of Nedd4L modulates its interaction with target substrates.^47,48 Thus, we investigated whether RSL3 modulates the interaction between Nedd4L and BMPR2. Co-IP revealed that the association between Nedd4L and BMPR2 was decreased in response to RSL3 treatment (Fig. 5f). These results indicate that lipid peroxidation induces the phosphorylation of Nedd4L, thus disassociating Nedd4L from BMPR2. To study the effect of Nedd4L phosphorylation, we examined the Nedd4L phosphorylation site by proteomics to investigate its mechanism. Mass spectrometry analysis indicated that threonine (T)451 is the possible phosphorylation site on Nedd4L regulated by RSL3 treatment (Fig. 5g). To identify the kinase responsible for T451 phosphorylation, we assessed the effect of BMPR2 on phosphorylation, since BMPR2 is a serine/threonine kinase that remains constitutively active. Overexpression of BMPR2 had no effect on the phosphorylation of Nedd4L (Supplementary Fig. 9a). We have revealed that Raf1 kinase exhibits a profibrotic role in lung fibrosis.⁴⁹ We observed that the Raf1 kinase inhibitor GW5074 attenuated RSL3-induced threonine phosphorylation of Nedd4L (Supplementary Fig. 9b). Furthermore, overexpression of Raf1 increased threonine phosphorylation of Nedd4L but not the threonine 451 mutant of Nedd4L (T451A) (Fig. 5h), indicating that Raf1 may modulate Nedd4L phosphorylation on threonine 451.

Fibroblast-specific Nedd4L deficiency exacerbates experimental PF

Our previous study demonstrated that upregulation of Nedd4L in the lungs mitigated the severity of experimental PF.³¹ To further assess the role of fibroblast-specific Nedd4L loss in PF pathogenesis, we created inducible fibroblast-targeted Nedd4L knockout mice (Fb-Nedd4L^−/−) through the cross of Nedd4L floxed mice (Nedd4L-LoxP) with tamoxifen-inducible Col1a2-Cre transgenic mice (Fig. 6a). Col1a2 is expressed in mesenchymal tissues, including lung fibroblasts, and tamoxifen-inducible Col1a2 promoter mice have been widely used to delete genes in lung fibroblasts. Fb-Nedd4L^−/− mice were confirmed by genotyping (Fig. 6b). Nedd4L mRNA levels was evaluated to be reduced in in lung tissues from Fb-Nedd4L^−/− mice relative to those from Nedd4L-LoxP (control) mice (Fig. 6c). To assess whether Nedd4L-mediated regulation of BMPR2 stability, fibroblasts isolated from Nedd4L-LoxP lungs were treated with control adenovirus or Cre-expressing adenovirus for 7 days. Nedd4L levels were markedly decreased in Cre-positive cells, correlating with a reduction in BMPR2 levels (Fig. 6d), suggesting that Nedd4L also stabilizes BMPR2 in mouse lung fibroblasts.

Fig. 6.

Nedd4L deficiency in fibroblasts exacerbates bleomycin-induced lung fibrosis. a Schematic of the design of Col1a2-Cre/Nedd4L-LoxP mouse generation. b DNA gel electrophoresis analysis of Col1a2-Cre/Nedd4L-LoxP and Nedd4L-LoxP mice by genotyping. c Real-time PCR was used to examine the Nedd4L mRNA levels in the lung tissues of Col1a2-Cre/Nedd4L-LoxP and Nedd4L-LoxP mice (n = 3). d Lung fibroblasts were isolated from lung tissues of Nedd4L-LoxP mice and then infected with adenovirus (control/Cre). BMPR2, Nedd4L and Cre protein levels were examined by immunoblotting (n = 2). e Schematic of the experimental design. f Survival rate (%) of bleomycin-challenged mice (n = 8–13). g H&E and Masson’s trichrome staining of lung tissues from bleomycin (2 weeks)-challenged mice (n = 4–6). Scale bar = 100 μm. h Ashcroft soring was analyzed based on Masson’s trichrome staining. i Real-time PCR was performed to analyze the mRNA expression of Col1A1 and fibronectin (FN) in the lung tissues of bleomycin (2 weeks)-challenged mice (n = 4–6). j Immunoblotting analysis of FN in lung tissue lysates. FN intensities were analyzed by ImageJ (n = 3). ^*p < 0.05, ^***p < 0.001

Furthermore, C57BL/6 wild-type mice, Nedd4L-LoxP, and Fb-Nedd4L^−/− mice were subjected to intratracheal (i.t.) bleomycin or PBS administration (Fig. 6e). Nedd4L levels in bleomycin-challenged Fb-Nedd4L^−/− mice were dramatically lower than those in Nedd4L-LoxP mice (Supplementary Fig. 10). Loss of Nedd4L in fibroblasts resulted in a mortality rate more than double that of Nedd4L-LoxP controls (Fig. 6f). Notably, the survival rate of Nedd4L-LoxP mice was less than that of wild-type mice (Fig. 6f), while no significant histological changes were observed in PBS-treated mice (Supplementary Fig. 11). It is possible that the floxed allele is inherently hypomorphic. H&E and Masson’s trichrome staining and Ashcroft scoring demonstrated that Fb-Nedd4L^−/− mice exhibited exacerbated bleomycin-induced lung fibrosis than control mice (Fig. 6g, h). There were significant increases in Col1A1 and fibronectin (FN) mRNA, as well as FN protein levels, in bleomycin-challenged Fb-Nedd4L^−/− mice (Fig. 6i, j). Furthermore, we found that Fb-Nedd4L^−/− mice exhibited an earlier onset of fibrosis than control mice (Supplementary Fig. 12a–d). The effect of Nedd4L deficiency in lung fibroblasts on LPS-induced postinjury lung fibrosis were determined (Supplementary Fig. 13a). Fb-Nedd4L^−/− mice delayed lung repair and increased fibrotic responses following LPS challenge compared to control mice (Supplementary Fig. 13b). LPS-induced TGF-β1-3 levels in the lungs were further elevated in Fb-Nedd4L^−/− mice (Supplementary Fig. 13c). Furthermore, we observed that compared to control wild-type (Col1a2-Cre-Nedd4L-loxP) mice, 13-month-old Fb-Nedd4L^−/− mice exhibited increased collagen levels in the lungs (Supplementary Fig. 14), suggesting that fibroblast-specific Nedd4L deficiency promotes spontaneous fibrosis. Together, these results support that loss of Nedd4L in lung fibroblasts facilitates the development of lung fibrosis.

Overexpression of BMPR2 diminishes bleomycin-induced lung fibrosis in Fb-Nedd4L^−/− mice

We assessed whether overexpression of BMPR2 rescues lung fibrosis in Fb-Nedd4L^−/− mice. Fb-Nedd4L^−/− mice received intratracheal administration of lentivirus-BMPR2 (Lenti-BMPR2) or control lentivirus (Lenti-Cont) 3 days prior to bleomycin challenge, followed by tissue collection on day 14 (Fig. 7a). Lenti-BMPR2 administration robustly upregulated BMPR2 at both the mRNA and protein levels (Fig. 7b, c) and increased the phosphorylation of Smad1/5 in mouse lung tissues (Fig. 7c), suggesting that lenti-BMPR2 administration activates BMP signaling in the lungs. The mortality rate, Col1A1 and FN mRNA expression, and FN protein levels were significantly reduced in BMPR2-overexpressing mice (Fig. 7d, e). Compared to Lenti-Cont mice, lung fibrosis and collagen deposition were reduced in lenti-BMPR2 mice (Fig. 7g, h). These data indicate that a protective role for BMPR2 against lung fibrosis in Fb-Nedd4L^−/− mice. Furthermore, we assessed the levels of TGFβ and related metabolites in BAL fluids. All TGFβ isoforms, including TGF-β1, TGF-β2, and TGF-β3, were elevated in bleomycin-challenged Nedd4L-Loxp mice and further increased in bleomycin-challenged Fb-Nedd4L^−/− mice. These increases were attenuated in Fb-Nedd4L^−/− mice overexpressing BMPR2 via Lenti-BMPR2 (Fig. 7i), indicating that BMPR2 overexpression reduces TGF-β production. Principal component analysis (PCA) revealed distinct clustering of the experimental groups (Fig. 7j). Striking alterations in multiple cellular metabolites were observed (Fig. 7k), including 3-hydroxy-2-methylbutanoic acid and 6-hydroxy-5-methoxylinodole glucuronide. These metabolites were increased in Nedd4L-Loxp mice and further elevated in Fb-Nedd4L^−/− mice, but their levels were attenuated in Fb-Nedd4L^−/− mice overexpressing BMPR2 (Fig. 7l).

Fig. 7.

Overexpression of BMPR2 mitigates experimental PF in fibrblast-Nedd4L-deficient mice. a Schematic of the experimental design. b Real-time PCR was performed to analyze the mRNA levels of BMPR2 in the lung tissues of lentivirus-control and lentivirus-BMPR2-infected mice (n = 3–4). c BMPR2 and P-Smad1/5 levels were analyzed by immunoblotting in lung tissues of lentivirus-control and lentivirus-BMPR2-infected mice (n = 3–4). The intensities of BMPR2 and P-Smad1/5 were examined by ImageJ. d Survival rate (%) of bleomycin-challenged mice (n = 7–9). e Real-time PCR was performed to examine the mRNA levels of Col1A1 and FN in lung tissues from mice subjected to experimental PF for 2 weeks (n = 4–6). f Immunoblotting analysis of FN in lung tissue lysates. FN intensities were analyzed by ImageJ (n = 3). g H&E and Masson’s trichrome staining of lung tissues from lentivirus-control and lentivirus-BMPR2-infected mice challenged with bleomycin (n = 4–6). Scale bar = 100 μm. h A hydroxyproline assay was performed to analyze the soluble collagen levels in mouse lung tissues (n = 3–5). ^*p < 0.05, ^**p < 0.01, ^***p < 0.001. i TGF-β1-3 levels in BAL fluids were measured by MSD multiplex assay (n = 4). ^*p < 0.05, ^**p < 0.01, ns: no significance. j Score plot of BAL samples using the PLS-DA model. k Unsupervised heatmaps of metabolites in high-flow heated electrospray ionization (HESI) positive mode and negative mode. l Examples of two metabolites in BAL fluids across different experimental groups

Discussion

BMP/BMPR2 signaling plays a protective, antifibrotic role in lung fibrosis. Previous studies have reported that reduced BMPR2 expression in lungs was observed in both patients with IPF and experimental PF models; however, the molecular mechanisms that regulate BMPR2 levels remain unexplored. In this study, we demonstrate that elevated lipid peroxidation in lung fibroblasts promotes BMPR2 degradation through the ubiquitination-mediated protein degradation system. We further revealed that the E3 ubiquitin ligase Nedd4L enhances the stability of BMPR2 by increasing its K63-linked polyubiquitination. Mechanistically, lipid peroxidation during the development of lung fibrosis increases phosphorylation of Nedd4L at threonine 451 and disrupts its interaction with BMPR2. Importantly, mice lacking Nedd4L in fibroblasts exhibit more severe lung fibrosis, which can be mitigated by overexpression of BMPR2. These findings reveal a new mechanistic insight into how Nedd4L maintains BMPR2 stability and suggest that enhancing BMPR2 levels may offer a promising therapeutic approach for PF.

Emerging evidence highlights the antifibrotic function of BMPs in PF. Guan et al. discovered that exogenous BMP4 reversed TGF-β1-induced myofibroblast differentiation and mitigated experimental PF.¹¹ Similarly, FK506 (tacrolimus), an immunosuppressant, indirectly activated the BMP pathway and reduced the severity of experimental murine PF.^50–52 BMPR2 is a key receptor for the BMP signaling pathway with constitutive kinase activity. Although the biological roles of BMPR2 have been extensively investigated in PAH,^18–20 its role in lung fibrosis remains unexplored.^16,17 Fukihara et al. reported that BMPR2 overexpression antagonized the TGF-β1 signaling pathway in lung fibroblasts.¹⁶ In our study, we show that BMPR2 overexpression significantly attenuates experimental lung fibrosis. This represents the first preclinical evidence directly demonstrating the antifibrotic function of BMPR2 in PF, underscoring its potential as a therapeutic target. Activation of BMPR2 may unintentionally worsen vascular remodeling, given that BMP activity has been shown to promote vascular calcification under conditions of hyperglycemia and diabetes.⁵³ Thus, broad upregulation of BMPR2 in diverse cellular populations could lead to unintended safety risks. Therefore, elevating fibroblast-specific BMPR2 levels may be a potential therapeutic strategy to treat fibrotic diseases, including IPF.

Recent findings have shown that BMPR2 levels are downregulated in lung tissues from patients with IPF, as well as those with both IPF and PAH.²¹ In this study, we found that BMPR2 undergoes degradation in fibrotic lungs. Although elevated lipid peroxidation has been found in IPF, its specific impact on fibrotic signaling, especially on the BMP pathway, has not been defined.^29,30 Our findings suggest that reductions in BMPR2 and BMP/BMPR2 signaling correlate with heightened lipid peroxidation and elevated oxidative stress throughout fibrotic progression in the lung.

Previous studies, including our own, have demonstrated that Nedd4L possesses antifibrotic properties by attenuating TGF-β signaling.^31,32 Because both BMPR2 and Nedd4L expression are reduced in lung fibrosis, we speculated that these proteins are functionally interconnected. We found that Nedd4L promotes stabilization of BMPR2 by enhancing its K63-linked polyubiquitination. Rather than the degradation-associated K48-linked polyubiquitination, K63-linked polyubiquitination is generally linked to protein stabilization and the modulation of signaling regulation.^40–43 Consistent with this, recent studies have demonstrated that K63-linked polyubiquitination stabilizes substrate proteins such as MCL-1, p53, AMOTL, and KLF4.^54–57 Our findings suggest that K63-linked polyubiquitination antagonizes K48-linked polyubiquitination. Specifically, Nedd4L enhances K63-linked polyubiquitination, thereby limiting the addition of K48-linked polyubiquitin chains to BMPR2. Ubc5B, an E2 ubiquitin-conjugating enzyme, facilitates the formation of K63-linked polyubiquitin chains for Nedd4L. This raises the possibility that Ubc5B contributes to BMPR2 stability and activation of Smad1/5 signaling. Smurf1, an E3 ubiquitin ligase, has been linked to the degradation of BMPR2 in mesenchymal stem cells,⁴⁴ but we found that it has no effect on BMPR2 degradation in lung fibroblasts. Our data showed that HECTD2 E3 ligase overexpression reduced BMPR2 and P-Smad1/5 levels (Supplementary Fig. 15a). HECTD2 targets and increases K48-linked polyubiquitination of BMPR2, which was attenuated by Nedd4L overexpression in lung fibroblasts (Supplementary Fig. 15b, c). However, this study did not focus on determining whether Nedd4L competes with HECTD2 for binding to the same PY motif. While no direct link between specific Nedd4L SNPs and IPF has been reported, rs513563, rs3865418, rs4149589, and rs3865418 have been shown to be associated with hypertension in US and Greek white cohorts.⁵⁸ Given that systemic hypertension is a significant risk factor and frequent comorbidity in IPF,⁵⁹ these variants could contribute to IPF susceptibility if they affect Nedd4L E3 ligase activity, at least among US white and Greek white populations. Thus, additional studies are required to test this hypothesis and to evaluate whether these Nedd4L genetic variants are associated with IPF.

Nedd4L activity is regulated through phosphorylation, with kinases such as Akt1, SGK1, and IKKβ known to modify its function via phosphorylation at specific residues.^60–62 For instance, phosphorylation at serine 448 has been shown to control Nedd4L self-ubiquitination and disrupt its interaction with ENaC.^48,60,63 In this study, we identified that lipid peroxidation promotes the phosphorylation of Nedd4L at threonine 451. This modification appears to reduce the interaction between Nedd4L and BMPR2. Interestingly, several functional phosphorylation sites on Nedd4L, including serine 488 and the newly identified threonine 451, are located between its WW domains, suggesting that serine/threonine phosphorylation may regulate substrate binding. Our findings indicate that Raf1 mediates threonine 451 phosphorylation of BMPR2. Future studies are warranted to define the mechanisms by which Raf1-mediated phosphorylation at threonine 451 regulates the Nedd4L-BMPR2 interaction.

In summary, this study reveals that Nedd4L plays a critical role in stabilizing BMPR2, thereby enhancing BMP/BMPR2-mediated antifibrotic signaling. Mechanistically, Nedd4L promotes K63-linked ubiquitination, which may counteract K48-linked ubiquitination of BMPR2. These findings uncover a novel molecular mechanism underlying PF progression and suggest that targeting Nedd4L and BMPR2 expression or activity in lung fibroblasts may represent a promising therapeutic strategy for IPF. Lentiviral BMPR2 is not an ideal therapeutic strategy due to significant limitations, including the risk of insertional mutagenesis, uncertain long-term safety, and potential immune responses. Agents such as FK506 and ataluren have been designed to restore BMPR2 signaling for the treatment of PAH, but their functional effects in lung fibrosis remain insufficiently evaluated. Our findings identify Nedd4L as a new therapeutic target that enhances BMPR2 signaling while concurrently suppressing TGFβR signaling. Future development of fibroblast-specific strategies to modulate Nedd4L and BMPR2 expression using nonintegrating viral vectors (such as adeno-associated virus) or mRNA-based therapies could offer a promising new approach for IPF.

Materials and methods

Cell culture and reagents

Human lung fibroblasts (IMR90 and Mrc5), human telomerase reverse transcriptase-introduced human lung fibroblasts (HLF/hTERT), and HeLa cells were purchased from the American Type Culture Collection (ATCC; Manassas, VA). Primary human lung fibroblasts from adult control human subjects (HLF) and patients with IPF (IPF-LF) were obtained from the Center for Organ Recovery and Education and Lung Transplantation at the University of Pittsburgh. Mouse lung fibroblasts were isolated from mouse lung tissues. Fibroblast cultures were grown in Eagle’s minimum essential (EMEM) medium supplemented with 10% fetal bovine serum (FBS; Corning, New York). HeLa cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco, Carlsbad, CA) containing 10% FBS. All cells were maintained at 37 °C in a humidified atmosphere containing 5% CO₂.

Plasmid and siRNA transfection

V5-tagged expression constructs for human Nedd4L, Raf1, and BMPR2 were generated by inserting the corresponding cDNAs into the pcDNA3.1/V5-His-Topo expression vector. The Nedd4L-Xpress plasmid was obtained as a gift from Christie Thomas (Addgene plasmid #83433). Threonine 451 to alanine mutation in the Nedd4L plasmid (Nedd4LT451A-Xpress) was generated by site-directed mutagenesis. An HA-tagged BMPR2 was inserted into the pcDNA3.1 expression vector. BMPR2-YA/KR- mutant HA-tagged plasmids were generated from the BMPR2-HA wild-type construct using a site-directed mutagenesis approach. All small interfering RNAs (siRNAs) were purchased from Sigma. Plasmid transfection was performed with Genject Plus or Polyjet transfection reagent (SignaGen Laboratories). Pepmute transfection reagent (SignaGen Laboratories) was used with siRNAs to downregulate target genes.

Immunoblotting, coimmunoprecipitation (Co-IP), and ubiquitination assay

Proteins were isolated from cultured cells and mouse lung tissues using lysis buffer. Equivalent amounts of total protein (15–20 µg) were separated by SDS‒PAGE and transferred onto nitrocellulose membranes, which were subsequently probed with appropriate primary and secondary antibodies. For Co-IP assays, cell lysates were incubated with primary antibodies overnight at 4 °C, followed by incubation with protein A/G beads for an additional 2 h at 4 °C. The beads were washed four times with PBS, and bound proteins were eluted using 2 × SDS sample buffer and heated at 95 °C for 5 min. Immunoprecipitated proteins were analyzed by immunoblotting. For the in vivo ubiquitin assay, cell pellets were lysed in 2% SDS buffer supplemented with ubiquitin aldehyde and N-ethylmaleimide (NEM), followed by boiling at 100 °C for 10 min prior to Co-IP analysis.

RNA extraction and analysis

Total RNA was isolated from cultured cells and mouse lung tissues using a commercial RNA extraction kit (IBI Scientific). Complementary DNA (cDNA) synthesis was performed using a reverse transcription kit from Bio-Rad. Quantitative real-time PCR was carried out using the SsoFast Eva Green Supermix (Bio-Rad). Primer sequences were designed based on corresponding human or mouse mRNA sequences.

Immunofluorescence staining

Human lung fibroblasts (Mrc5 and hLF/HERT) were cultured on glass-bottom dishes. After the indicated treatments, the cells were fixed with 4% paraformaldehyde and blocked with 5% BSA (bovine serum albumin) in TBST (Tris-buffered saline with Tween 20) buffer. Fixed cells were subsequently incublated with fluorescence-conjugated antibodies and DAPI (4’,6-diamidino-2-phenylindole) were incubated with fixed cells. Immunofluorescent cell imaging was performed using a confocal microscope (Nikon). BMPR2 fluorescence signals were quantified by analyzing the equivalent square areas of the plasma membrane and cytoplasm using ImageJ software.

Reanalysis of publicly available database

Deposited RNA-seq datasets (GDS4279, GDS4580, GDS4902, and GDS5078) in the Gene Expression Omnibus (ncbi.nim.nih.gov/geo) were downloaded. The data were normalized to either healthy subjects or PBS-treated mice to calculate fold changes in BMPR2 expression. Gene expression of IPF lungs was obtained from the Lung Genomics Research Consortium (LGRC) gene expression dataset (http://www.lung-genomics.org). Correlation analysis was performed using simple linear regression analysis with Prism 10 (GraphPad, San Diego, CA).

Animal models and flow cytometry

C57BL/6, Nedd4L-LoxP, and Col1a2-Cre/Nedd4L-LoxP mice were maintained under a specific pathogen-free condition in the animal care facility at The Ohio State University. All animal procedures (2018A00000122-R2) were approved by the Institutional Animal Care & Use Committee (IACUC) of The Ohio State University. For the bleomycin model, all mice (male, aged 8–10 weeks) were intratracheally administered with bleomycin (1.2 or 2.2 µ/kg) and euthanized at time points ranging from 1 to 5 weeks post-treatment. For the repetitive bleomycin model, C57BL/6 mice were challenged with intratracheal bleomycin (1.2 µ/kg) six instillations at 2-week intervals. For the LPS model, all mice were intratracheally administered LPS (E. coli. O127:B8, 5 mg/kg) for 7 days. For tamoxifen injection, mice were intraperitoneally injected with tamoxifen by single daily injections for five consecutive days. Lentivirus (1 × 10⁸ TU/mL) was mixed with Lipofectamine 2000 reagent (Invitrogen), followed by intratracheally injection into the mouse lungs. For the flow cytometry assay, lung fibroblasts were isolated from mice that had been injected with bleomycin for 2 weeks. Ashcroft scoring was used to evaluate the extent of pulmonary fibrosis using a numerical scale from 0 to 8. TGF-β1-3 levels in BAL fluids were measured using multiplex cytokine analysis (Meso Scale Discovery) according to the manufacturer’s instructions.

Metabolomics analysis

BAL fluid (200 μl) from each sample was extracted with 800 μl methanol, vortexed, centrifuged, dried, and then reconstituted in 400 μl of 80% methanol. Finally, 100 μl of the supernatant was transferred to a 2 ml HPLC vial for LC‒MS/MS analysis. In addition, quality control was prepared by pooling 20 μl from the supernatant of each sample. Untargeted metabolomics analysis was performed on a Vanquish UHPLC coupled to an Orbitrap Exploris 480 mass spectrometer (Thermo Fisher Scientific, Waltham, MA). The metabolites were separated on an Accucore C18 2.1 × 100 mm, 2.6 μm column (Thermo Fisher, USA) at a flow rate of 0.5 mL/min. Binary mobile A (0.1% formic acid in nanopure water) and B (0.1% formic acid in 95% acetonitrile in water (v:v)) were used with the following gradient: 0–0.5 min, hold at 5% B; 0.5–2 min, 5% to 25% B; 2–6 min, 25% to 35% B; 6–9 min, 35% to 100% B; 9–11 min, hold at 100% B; 11.1–13 min, equilibrate at 5% B. The column temperature was set at 40 °C. An injection volume of 5 μl was used during the analysis. All the samples were analyzed in a random order. The quality control sample and processing blank were injected at the beginning of the sequence and after every 10 sample injections. Each sample was analyzed in 3 technical replicates. The HESI positive mode and negative data were collected separately. Compound Discoverer (v 3.4m Thermo Fisher Scientific, USA) was utilized to perform data processing, including peak picking, retention time alignment, background removal, ion extraction and integration, and putative identity annotation. Multivariate data analysis, including principal component analysis (PCA) and heatmaps, was performed using the open-source MetaboAnalyst 6.0.

Quantification and statistical analysis

Band intensities were quantified and analyzed using either Student’s t test or analysis of variance (ANOVA), as appropriate. Data are presented as the mean ± SD of samples from at least three independent experiments. A p value < 0.05 was considered statistically significant.

Author contributions

Q.Y., B.B., N.S., S.J.T., G.C., N.L.P., Y.Z., M.R., C.G., Q.M., J.H., J.Z., Y.Z.: methodology, investigation, formal analysis, validation, and visualization. Q.Y.: conceptualization, writing-orignial draft, funding acquisition. Y.Z.: conceptualization, supervision, resources, funding acquisition, and writing-review and editing. S.J.T.: editing. J.Z., D.J.K., and M.R.: resources and editing. All the authors have read the journal’s authorship agreement. All authors have read and approved the article.

Data availability

The raw metabolic analysis data reported in this paper have been deposited in the Metabolomics Workbench, an international repository for metabolomics data and metadata, metabolite standards, protocols, tutorials and analysis tools. The DOI for this project (PR002952) is: https://doi.org/10.21228/M83C44.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41392-026-02786-2.