Persistence of alveolar fibroblast-derived ADAMTS4+ cells in a preclinical model of delayed pulmonary fibrosis resolution

Zabihi, Mahsa; Khadim, Ali; Lingampally, Arun; Vazquez-Armendariz, Ana Ivonne; Hadzic, Stefan; Panagiotidis, Georgios-Dimitrios; Kalina, Daniel; Halweg, Jan; Procida-Kowalski, Tara; Bartkuhn, Marek; Chu, Xuran; Koepke, Janine; Samakovlis, Christos; Boehm, Mario; Weissmann, Norbert; Günther, Andreas; Seeger, Werner; Braubach, Peter; Herold, Susanne; Wygrecka, Malgorzata; Bellusci, Saverio; El Agha, Elie

doi:10.1038/s41467-026-72419-3

Download PDF

Article
Open access
Published: 08 May 2026

Persistence of alveolar fibroblast-derived ADAMTS4+ cells in a preclinical model of delayed pulmonary fibrosis resolution

Nature Communications volume 17, Article number: 4205 (2026) Cite this article

2044 Accesses
5 Altmetric
Metrics details

Subjects

Abstract

Idiopathic pulmonary fibrosis is the most common and aggressive form of interstitial lung disease. Despite extensive research on the pathomechanisms of fibrogenesis, little is known about the mechanisms of fibrosis resolution. Here, lineage tracing of alveolar fibroblasts was carried out during fibrosis development and delayed resolution in aged mice. Histological analyses, single-cell transcriptomics, and ex vivo models including alveolar organoids and precision-cut lung slice cultures were employed. The data reveal that lipofibroblasts contribute to myofibroblast formation during fibrogenesis, with the reverse differentiation trajectory occurring during fibrosis resolution. Importantly, delayed resolution is associated with the persistence of ADAM metallopeptidase with thrombospondin type 1 motif 4-positive (ADAMTS4+) cells. Investigation of human lung transplant tissues, single-cell and spatial transcriptomic datasets, and functional ex vivo interventions reveal strong clinical relevance. Our study underscores the significance of the lipofibroblast-to-myofibroblast reversible switch in fibrosis development and resolution and identifies ADAM metallopeptidase with thrombospondin type 1 motif 4 as a potential therapeutic target in human lung fibrosis.

AAV9-Tspyl2 gene therapy retards bleomycin-induced pulmonary fibrosis by modulating downstream TGF-β signaling in mice

Article Open access 30 June 2023

Autocrine TGF-β-positive feedback in profibrotic AT2-lineage cells plays a crucial role in non-inflammatory lung fibrogenesis

Article Open access 31 August 2023

Alveolar fibroblast lineage orchestrates lung inflammation and fibrosis

Article 10 July 2024

Introduction

Idiopathic pulmonary fibrosis (IPF) is widely perceived as an aberrant wound-healing response resulting from repetitive microinjuries to the alveolar epithelium and continuous scarring of the alveolar network. This pathological course features excessive deposition of extracellular matrix (ECM) proteins and culminates in progressive loss of gas exchange and ultimately respiratory failure. IPF is the most common and relentless form of interstitial lung disease (ILD), with a median survival of 3–5 years after diagnosis^1,2,3. Aging, environmental exposures, and genetic factors are associated with an increased risk of developing IPF^4,5,6. This disease typically manifests in middle-aged adults, suggesting a strong link between aging and increased incidence and prevalence⁷. Hallmarks of aging including genomic instability, telomere shortening, cellular senescence, mitochondrial dysfunction, and elevated reactive oxygen species (ROS) have been implicated in IPF pathogenesis⁸. Notably, NADPH oxidase 4 (NOX4), an enzyme involved in ROS generation and regulation, is elevated in IPF lungs, and its inhibition has been shown to attenuate lung fibrosis⁹. Metformin, a well-known antidiabetic drug, has demonstrated antifibrotic effects via adenosine monophosphate-activated protein kinase (AMPK) signaling that inhibits NOX4¹⁰ and induces apoptosis in myofibroblasts (MyoFBs)¹¹, the effector cells that promote fibrotic remodeling in the lung^12,13,14. We previously showed that metformin inhibits collagen production in MyoFBs in an AMPK-dependent manner and induces lipogenic differentiation in an AMPK-independent manner¹⁵.

Lipofibroblasts (LIFs) are alveolar fibroblasts located adjacent to alveolar type 2 (AT2) cells¹⁶ These cells contain neutral lipid droplets and are characterized by the expression of adipose differentiation-related protein (Adrp a.k.a. perilipin 2 or Plin2)¹⁷, transcription factor 21 (Tcf21)¹⁸, and fibroblast growth factor 10 (Fgf10)^19,20. It is important to mention that there is no single marker or transgenic line that can currently be used to exclusively label or target LIFs. We previously lineage-traced Plin2+ cells in young adult mice to show that upon bleomycin injury, these cells give rise to MyoFBs²¹. Interestingly during fibrosis resolution in young mice, a subset of MyoFBs can revert to the native LIF phenotype, restoring normal lung architecture^15,21. However, this reversible switch is not seen in human IPF, leading to a progressive and irreversible state.

Further refinement of the LIF population enriched with stem cell antigen 1 (Sca-1)/Fgf10 double-positive cells suggested a higher potential for this subpopulation to support AT2 growth in alveolar organoids²². However, aging impairs the niche activity of resident mesenchymal cells (rMCs) due to increased Nox4 expression, which can be reversed by NOX4 inhibition²³. Moreover, lipid scarcity in aging lungs has been implicated in reduced AT2 renewal capacity, and lipid replenishment has been shown to restore this activity in alveolar organoids²⁴.

Recent single-cell RNA sequencing (scRNA-seq) studies have highlighted the role of LIF-to-MyoFB differentiation during fibrosis formation^25,26,27,28. These studies identified transitional cell states, such as secreted frizzled related protein 1-positive (Sfrp1 + ), RUNX family transcription factor 1-positive (Runx1 + ), lipocalin 2-positive (Lcn2 + ), and alveolar fibroblast 3 (AL3) subpopulations that link LIFs and collagen triple helix repeat containing 1 (Cthrc1 + ) MyoFBs^27,29,30. Targeting these transitional states led to lung simplification and emphysema-like phenotypes thus emphasizing their importance in therapeutic strategies³⁰. However, most of these studies focused on young mice, which is a limitation since bleomycin-induced pulmonary fibrosis in young mice spontaneously resolves, failing to replicate the persistent fibrosis observed in human IPF patients. Since IPF is an age-related disease, studying fibrosis in aged mice may provide more clinically relevant insights. Aged mice exhibit delayed fibrosis resolution and prolonged transitional states, making them a better model for understanding differentiation trajectories during both fibrosis formation and delayed resolution^31,32.

In this study, we used 52–to-56-week-old female Fgf10^Cre-ERT2/+; tdTomato^flox mice to model fibrosis with delayed resolution, mimicking human IPF pathology. We performed scRNA-seq on lineage-labeled Fgf10+ cells following bleomycin-induced lung injury to investigate cellular heterogeneity and fate trajectories. Additionally, we used alveolar organoid and precision-cut lung slice (PCLS) cultures to test potential therapeutic interventions. We finally validated our findings by investigating publicly available databases and employing fresh human lung tissues.

Results

Aged mice demonstrate incomplete fibrosis resolution following bleomycin injury

Female Fgf10^Cre-ERT2/+; tdTomato^flox mice at 52-to-56 weeks of age were administered three intraperitoneal (i.p.) injections of tamoxifen followed by a single i.t. instillation of bleomycin (Bleo) or saline (Sal) (Fig. 1a). This experimental approach enabled the labeling of pre-existing Fgf10 + cells, allowing to track their fate during fibrosis formation and resolution. Lungs were harvested at day 14 (d14) (corresponding to the peak of fibrosis) and days 30 (d30) and 60 (d60) (corresponding to early and late fibrosis resolution, respectively) after the Bleo challenge (Fig. 1a). Hematoxylin and eosin (H&E) staining confirmed the development of pulmonary fibrosis by d14 where 40% ± 3.8% of the tissue was fibrotic (Supplementary Fig. 1a–c). Despite structural improvements from Bleo d30 till Bleo d60, persistent damage was still observed at these timepoints (24% ± 2.08% and 11.3% ± 1.4%, respectively) (Supplementary Fig. 1a–c). Masson’s Trichrome and Picro Sirius Red staining showed increased collagen deposition at Bleo d14, and to a lesser extent at Bleo d30 and Bleo d60 (Supplementary Fig. 1d–g).

Fig. 1: Characterization of lipofibroblast marker expression in normal and fibrotic aged Fgf10Cre-ERT2; tdTomatoflox lungs. — **Fig. 1: Characterization of lipofibroblast marker expression in normal and fibrotic aged *Fgf10*^Cre-ERT2; tdTomato^flox lungs.**

Interconversion of Fgf10+ cells between the LIF and MyoFB states during fibrosis formation and resolution

We then assessed the status of LIF markers in Sal and Bleo lungs using RNAscope combined with immunofluorescence for surfactant protein C (SFTPC) (Fig. 1). In Sal lungs, 56.1% ± 6.3% of tdTom+ cells also expressed Tcf21 and these cells were observed adjacent to SFTPC+ cells (AT2s) (Fig. 1b, k, l). Following bleomycin injury (Bleo d14), tdTom/Tcf21 double positive cells and AT2s were dramatically lost in fibrotic regions (Fig. 1c, j–l). Similarly, tdTom/Inmt/Robo2 triple-positive cells were decreased from 34.1% ± 2.6% in Sal to 7.09% ± 0.9% in Bleo d14 (Fig. 1f, g, m). These results suggest that tdTom+ cells lose their LIF profile at the peak of fibrosis. Conversely, during early (Bleo d30) and late (Bleo d60) fibrosis resolution, the number of AT2s and the expression of LIF markers in lineage-labeled cells were largely, yet incompletely, restored (Fig. 1d, e, h, i).

To characterize lineage-labeled Fgf10 + (tdTom + ) cells and precisely determine their fate during fibrosis development and resolution, live, mesenchymal tdTom+ cells were FACS-sorted from Sal, Bleo d14, Bleo d30, and Bleo d60 lungs (Fig. 2a, b). FACS analysis demonstrated significant abundance of tdTom+ cells in Bleo d14 lungs and sustained presence of these cells during resolution (Bleo d30 and d60) (Fig. 2c). Next, sorted tdTom+ cells were loaded for scRNA-seq. Integrative UMAPs showed that tdTom+ cells in Sal predominantly consist of alveolar fibroblasts 1 (AF1 or LIF; around 55%) and AF2 (around 36%) along with peribronchial fibroblasts (PeriFBs; around 6.6%) and traces of myofibroblasts (MyoFBs; around 2.6%) (Fig. 2d, e). The top DEGs for each cluster are shown in Fig. 2f. In situ hybridization on Sal lung sections confirmed such heterogeneity at steady state and showed that tdTom+ cells contain (Pi16 + ) AF2, and (Hhip + ) PeriFBs (Supplementary Fig. 2a–c) in addition to AF1s (Fig. 1). Importantly, the scRNA-seq analysis also showed that the MyoFB cluster peaked at Bleo d14 (7.3% of total cells) with partial regression during early (5.3% at Bleo d30) and late (3.9% at Bleo d60) resolution (Fig. 2d, e). Interestingly, the AF1 cluster expanded at the expense of AF2 during resolution (Fig. 2d, e).

Fig. 2: Single-cell transcriptomics reveals the heterogeneity of the Fgf10+ lineage in normal and fibrotic aged lungs. — **Fig. 2: Single-cell transcriptomics reveals the heterogeneity of the *Fgf10*+ lineage in normal and fibrotic aged lungs.**

To gain insights into differentiation trajectories, RNA velocity was carried out, and the results showed a clear transition from LIFs (AF1s) to MyoFBs at Bleo d14 and to a lesser extent at d30, while the reverse trajectory was observed at Bleo d60 (Fig. 2g). We then performed RNA velocity analysis with latent inferred time to further investigate differentiation trajectories, and the results showed a trajectory from a transitional state (Sfrp1 + ) to MyoFB/proliferative state and then LIFs (Fig. 2h). Notably, ADAM metallopeptidase with thrombospondin type 1 motif 4 (Adamts4) was among the top genes defining the terminal fate (Fig. 2h). Its expression was mainly observed in AF1s rather than MyoFBs, being induced at Bleo d14 and persisting at Bleo d60 (Fig. 2i–k and Supplementary Fig. 2d, e). ADAMTS4 was previously shown to mediate exuberant fibroblast activity leading to exaggerated immunopathology and lung damage after influenza virus infection³³ with one of its major substrates being versican (VCAN), which is an important immunomodulator that guides cell proliferation, migration, and differentiation during inflammation. ADAMTS4 is involved in maintaining neutrophil infiltration and therefore sustaining an inflammatory milieu in the lung. We therefore hypothesized that the ADAMTS4/VCAN axis is involved in delayed resolution of pulmonary fibrosis.

In situ hybridization and immunofluorescence were carried out and the data revealed that tdTom+ cells were accumulated in fibrotic regions of Bleo d14 lungs (35.1% ± 2.7%) compared with normal regions of Sal lungs (19.77% ± 1.02%) (Fig. 3a, b, i). While the majority of tdTom+ cells were enriched with Fgf10 expression in Sal (60.77% ± 3.78%) (Fig. 3a, k), only 6.3% ± 2.15% of tdTom+ cells expressed Cthrc1 (Fig. 3e, o). At the peak of injury, tdTom+ cells lose Fgf10 expression (Fig. 3b, k) and express Cthrc1 (MyoFB marker) (37.1% ± 2%) (Fig. 3f, o). With the start of resolution, tdTom+ cells start reacquiring Fgf10 expression and by Bleo d60, 38.8% ± 1.04% of them express Fgf10 (Fig. 3c, d, k). In contrast, only 5.7% ± 1.24% Cthrc1/tdTom double-positive cells were detected in the damaged regions of Bleo d60 (Fig. 3h, o). Further comparison of AF1s with MyoFBs revealed Gsn, Inmt, C3, Igfbp6, Macf1 and Clec3b as top DEGs in AF1s and Nrep, Postn, C1qtnf6 and Ppic in MyoFBs (Supplementary Fig. 2f). A heatmap showing bulk analysis of top DEGs in Sal, Bleo d14, Bleo d30, and Bleo d60 is shown in Supplementary Fig. 2g. The data showed a general switch to MyoFBs during fibrogenesis and to LIFs during resolution. The AF1 signature was highest in Bleo d60, which is in line with strong lipofibroblast differentiation and reinforcement during resolution (Supplementary Fig. 2g and Fig. Supplementary Fig. 3a). Bleo d14 was dominated by MyoFB and Bleo d30 showed gradual downregulation of MyoFB markers and induction of PeriFB and AF2 markers (Supplementary Figs. 2g and 3b). We also investigated the expression pattern of Adamts4 by RNAscope and found that it is indeed upregulated in tdTom+ cells at d14 and persists at d60 (Fig. 3e–h, l, m, q). Its expression showed minimal overlap with Cthrc1 (Fig. 3e–h, p), supporting the finding that Adamts4 is largely expressed in AFs rather than MyoFBs during delayed resolution (Fig. 2i–k).

**Fig. 3: Contribution of FGF10+ cells to myofibroblasts during fibrogenesis and persistent ADAMTS4+ cells during delayed resolution.**

We also subclustered the MyoFB cluster to gain insights into the heterogeneity of these pathological cells and the transcriptomic changes occurring during fibrogenesis and recovery (Supplementary Fig. 4a, b). The analysis revealed seven subclusters including Cthrc1^high, Cthrc1^low, Gli1^high, Prolif^high, LIF^high, AF2^high, and a very minor SMC subcluster (Supplementary Fig. 4b, c). This was confirmed by applying a MyoFB (Supplementary Fig. 4d) and LIF scores (Supplementary Fig. 4e) reflecting the expression of the top 485 DEGs (Source Data file) for the respective populations. Analyzing individual timepoints showed the presence of almost only LIF^high and Cthrc1^low MyoFBs in Sal, while Cthrc1^high and Gli1^high MyoFBs were predominant in Bleo d14 before regressing through resolution (Supplementary Fig. 4f). The AF2^high subcluster was prominent in Bleo d30 (Supplementary Fig. 4f). In Bleo d60, the LIF^high cluster represented the majority of the remaining cells (Supplementary Fig. 4f), confirming our previous finding that MyoFB-to-LIF transition marks fibrosis resolution^15,21,34. A heatmap of the top regulated genes in these subclusters is shown in Supplementary Fig. 4g. A similar analysis was conducted on the AF1 cluster and the analysis revealed ten subclusters including two LIF^high, LIF^Int, two LIF^low, AF2^high, AF2^low, LIF^Adamts4, Apoptotic fibroblasts (ApoptoticF), and Proliferative fibroblasts (ProlifF) (Supplementary Fig. 4h–m). The two LIF^high subclusters correspond to steady-state LIFs (red) and injury-induced LIFs (green). Injury-induced LIFs are enriched for regulation of apoptosis in response to DNA damage in the saline condition and programs such as ECM organization and remodeling and mesenchyme development during active fibrosis and resolution. Overall, the emerging picture from analyzing the AF1 cluster is that the LIF signature is dominant in Sal, lost during fibrogenesis, and re-emerges during late resolution (Bleo d60). Similarly to MyoFB subclustering, an AF2 signature emerges during early resolution (Bleo d30) (Supplementary Fig. 4h, i, l). A heatmap of the top regulated genes in these subclusters is shown in Supplementary Fig. 4m.

Inhibition of ADAMTS4 attenuates fibrogenesis ex vivo

To show a proof of concept that ADAMTS4 represents a potential therapeutic target and that its inhibition promotes fibrosis resolution, we carried out alveolar organoid assays by co-culturing WT alveolar epithelial cells (AECs) with Sca1 + CD31/CD45/EpCAM triple-negative resident mesenchymal cells (rMCs) according to our previously published protocol³⁵. Following sixteen days of culture, organoids were treated with recombinant transforming growth factor beta 1 (rTGFβ1) or vehicle (Veh) for three days followed by treatment with Veh or recombinant tissue inhibitor of metalloproteinase 3 (rTIMP-3) for three days (Fig. 4a). TIMP-3 is a known inhibitor of ADAMTS4 activity^36,37 although it also targets other metalloproteinases. Immunofluorescence and qPCR analysis showed that rTIMP3 treatment preserved organoid morphology (Fig. 4b–g), attenuated Col1a1, Acta2, and Cthrc1 expression (Fig. 4h–j), and restored the expression of Sftpc (AT2 marker) and Ager (AT1 marker) (Fig. 4 m, n). Of note, Adamts4 expression was induced by rTGFβ1, which was largely attenuated by rTIMP-3 treatment (Fig. 4k). The expression of endogenous Timp3 was largely unchanged by the treatments (Fig. 4l).

**Fig. 4: Inhibiting ADAMTS4 mitigates fibrogenesis in murine alveolar organoids.**

As another layer of validation, we also generated PCLS from WT mice and subjected them to vehicle or rTGFβ1 treatment for five days. Since fibrogenesis is reversible in the PCLS model if rTGFβ1 is removed from the culture medium, PCLS were further treated with rTGFβ1 in the presence of vehicle or rTIMP-3 (Fig. 5a). Similarly to the organoid experiments, rTIMP3 treatment attenuated rTGFβ1-mediated fibrogenesis (Fig. 5b–d). Moreover, Adamts4 per se was also downregulated upon rTIMP-3 treatment (Fig. 5e) while Timp3 did not show significant changes (Fig. 5f). To add a functional dimension to these observations, protein lysates were subjected to an ADAMTS4 activity assay, and the results showed a significant increase in ADAMTS4 activity upon rTGFβ1 treatment, which was rescued by rTIMP-3 (Fig. 5g). To investigate the expression of ADAMTS4 in various mesenchymal subsets, thin sections were generated from the treated PCLS groups and RNAscope (for top markers of mesenchymal clusters) and immunofluorescence (for ADAMTS4) were carried out. The analysis confirmed the upregulation of ADAMTS4 at the protein level by rTGFβ1 treatment, and subsequent downregulation by rTIMP-3 (Fig. 5h–q). Moreover, the results showed that rTGFβ1 downregulates Tcf21 (AF1 marker) (Fig. 5r) while inducing Pi16 (AF2) (Fig. 5s), Cthrc1 (MyoFB) (Fig. 5t), and to some extent Hhip (PeriFB) (Fig. 5u), which was largely reversed by rTIMP-3 treatment. Recombinant TGFβ1 induced the expression of ADAMTS4 mostly in AF1s (Fig. 5r) and AF2s (Fig. 5s), which was reversed by rTIMP-3, especially in AF1s. In another set of experiments, PCLS from aged Bleo d14 Fgf10^Cre-ERT2; tdTomato^flox mice were cultured ex vivo for three days while being treated with recombinant VCAN (rVCAN) or Veh, and analysis by qPCR, in situ hybridization, and immunofluorescence also showed attenuation of fibrosis and ADAMTS4 expression (Supplementary Fig. 5).

**Fig. 5: Inhibiting ADAMTS4 mitigates fibrogenesis in murine precision-cut lung slice cultures.**

To validate our findings in the human context, we obtained lung explant tissues from IPF and non-IPF donors and carried out qPCR. The results showed significant upregulation of the MyoFB-associated genes COL1A1, ACTA2, CTHRC1, and SPP1 (Fig. 6a) and downregulation of SFTPC and HOPX (as expected) (Fig. 6b), but also significant downregulation of AF1-associated genes PDGFRA, CEBPB, LIMCH1, TCF21, INMT, PLIN2, PPARG, and NPNT (Fig. 6c). We also observed downregulation of the endothelial marker PECAM1 (Fig. 6d) and an upward trend for ADAMTS4 (Fig. 6e). We also carried out RNAscope for CTHRC1 and ADAMTS4 on formalin-fixed, paraffin-embedded donor and IPF lung tissues and the data showed an increase in the abundance of CTHRC1 + , ADAMTS4+, and double-positive cells in IPF lungs in parallel to a decrease in the abundance of cells expressing the AF1 markers INMT and ROBO2 (Fig. 6f–k). We also analyzed the human lung cell atlas (HLCA) by filtering for the stromal fraction in healthy, IPF, ILD, and non-specific interstitial pneumonia (NSIP) while preserving the original cluster annotation (Fig. 7a). Of note, cluster frequency analysis revealed significant loss of the AF1 cluster in IPF, ILD, and NSIP compared with the healthy state (Fig. 7b). We also carried out GSEA, and the results showed significant downregulation of the AF1 signature (defined in the Source Data file) in all fibrotic disease states (Fig. 7c). UMAP plots also showed downregulation of AF1 markers FGF10, SCUBE2, INMT, and LIMCH1 in parallel to upregulation of CTHRC1 and ADAMTS4 (Fig. 7d). Importantly, interrogating the top 10 DEGs in an unbiased manner revealed that ADAMTS4 is among the top genes that are upregulated in NSIP, IPF, and ILD compared with the healthy state (Fig. 7e). We also generated a volcano plot comparing IPF versus healthy, and ADAMTS4 was among the top upregulated genes in IPF (Fig. 7f).

**Fig. 6: Characterization of lipogenic and myogenic markers in human IPF lungs.**

Fig. 7: Increased expression of ADAMTS4 in fibrotic human lung disease. — **Fig. 7: Increased expression of *ADAMTS4* in fibrotic human lung disease.**

We then leveraged a published spatial transcriptomic dataset (S-BSST1410)³⁸ to interrogate the abundance and spatial distribution of various cell populations in donor and IPF (Fig. 8). As expected, we found significant loss of alveolar epithelial cells and accumulation of MyoFBs in IPF as indicated by downregulation of marker genes such as NAPSA, SFTPC, AGER, and HOPX in parallel to upregulation of genes related to collagen production and ECM remodeling (Fig. 8a, b). We also observed an overrepresentation of airway epithelial cells in IPF, indicating bronchiolization (Fig. 8a). Notably, TIMP3 was among the top downregulated genes in IPF compared with donor (Fig. 8b). Spatial visualization of CTHRC1 expression showed significant upregulation in the IPF samples, where its expression pattern mimicked that of MyoFBs (Fig. 8c vs. Fig. 8a). Importantly, ADAMTS4 was also significantly upregulated in IPF samples (Fig. 8d, e), and there was a positive correlation between spots expressing ADAMTS4 and those expressing CTHRC1 in IPF (Fig. 8f), further confirming that ADAMTS4 is upregulated in IPF and that its expression correlates with fibrogenesis. Collectively, the human single-cell and spatial transcriptomic data indicate that our findings using mouse-derived samples potentially possess strong clinical relevance in humans with pulmonary fibrosis.

Fig. 8: Spatial transcriptomics reveals upregulation of ADAMTS4 in human IPF. — **Fig. 8: Spatial transcriptomics reveals upregulation of *ADAMTS4* in human IPF.**

Precision-cut lung slices were also generated from fresh IPF lung explants and treated with rTIMP-3 or rVCAN for three days (Fig. 9a). Gene expression analysis revealed a significant downregulation of COL1A1 expression upon treatment (Fig. 9b). The treatment also strongly downregulated ADAMTS4 (Fig. 9b) similarly to what was observed with mouse-derived samples (Fig. 5e and Supplementary Fig. 5b). Further histological analysis using in situ hybridization showed attenuation of CTHRC1 and ADAMTS4 in response to rTIMP-3 or rVCAN treatment while restoring ROBO2 (AF1 marker) expression (Fig. 9c–f). Since rTIMP-3 is not a selective inhibitor of ADAMTS4 activity, siRNA-mediated knockdown of ADAMTS4 was also carried out. Human PCLS derived from non-IPF donors were treated with Veh or rTGFβ1 for five days followed by similar treatment in the presence of scrambled siRNA or siRNA targeting ADAMTS4 for three days (Fig. 9g). Immunofluorescence showed upregulation of ADAMTS4 at the protein level in response to rTGFβ1 treatment, which was attenuated by treatment with si-ADAMTS4 (Fig. 9h). As another layer of validation, qPCR and western blotting were performed and the results showed significant upregulation of COL1 and an upward trend for ADAMTS4 in response to rTGFβ1 treatment, which was attenuated by ADAMTS4 gene silencing (Fig. 9i–k). Further analysis confirmed that ADAMTS4 activity was enhanced by rTGFβ1 treatment, which was attenuated in si-ADAMTS4-treated samples (Fig. 9l). Finally, RNAscope confirmed that ADAMTS4 and CTHRC1 were upregulated by rTGFβ1 treatment, which was attenuated by si-ADAMTS4 treatment (Supplementary Fig. 6a–f, h). Knockdown of ADAMTS4 also restored ROBO2 expression (Supplementary Fig. 6b–e, g). Collectively, the data suggest that ADAMTS4 represents a potential therapeutic target in human lung fibrosis.

Fig. 9: ADAMTS4 loss of function attenuates fibrogenesis in human precision-cut lung slice cultures. — **Fig. 9: *ADAMTS4* loss of function attenuates fibrogenesis in human precision-cut lung slice cultures.**

Discussion

Ever since lipid-droplet-containing fibroblasts were reported in the alveolar interstitium of neonatal rat lungs³⁹, there has been controversy regarding the function or even existence of these elusive cells, which were later dubbed “LIFs”. Apart from their involvement in assisting AT2s with pulmonary surfactant production, LIFs are widely regarded as niche cells for AT2s⁴⁰ as they produce important growth factors such as FGF10 that is important for AT2 self-renewal and maintenance^41,42. Of note, while LIFs were sometimes claimed to only exist in rodent lungs⁴³, their presence was later confirmed in the human lung⁴⁴. Despite that, the term “LIFs” is still often ignored in single-cell atlases of the healthy and fibrotic lung where the corresponding cells are generally assigned the AF label. Earlier work had demonstrated that fibroblast-derived radiolabeled lipids that are encapsulated by PLIN2 are taken up by AT2s¹⁷. Due to their strong expression of Plin2, the latter had been used as a marker of LIFs. We previously demonstrated that LIFs are a source of MyoFBs in the fibrotic lungs of young mice²¹. In that study, we lineage-traced Plin2+ cells in the context of bleomycin-induced pulmonary fibrosis and subsequent resolution. However, the advent of single-cell transcriptomics allowed the identification of more reliable markers of LIFs thus enabling to carry out lineage tracing and the downstream analysis in a more precise manner. FGF10 is a reliable marker of AF1s as evident from single-cell atlases and multiple other studies, and the LIF signature is enriched in the AF1 fraction of AFs.

Against this background, we utilized our Fgf10^Cre-ERT2 knock-in line⁴⁵ to lineage-trace pre-existing Fgf10+ cells in the context of pulmonary fibrosis. Since bleomycin-induced pulmonary fibrosis spontaneously resolves in young mice while human IPF is progressive/irreversible, we utilized old mice in the current study to create a state of delayed fibrosis resolution that we believe better mimics the clinical setting. Our data clearly demonstrate that the LIF^high fraction of Fgf10+ cells contributes to MyoFB formation during fibrosis development. These data are not only in line with our previous work using young mice²¹, but also with more recent literature implicating AFs as a major source of MyoFBs in the fibrotic lung^25,26,27,34. During resolution, MyoFBs undergo apoptotic clearance or deactivation. The conversion of MyoFBs into AFs with a LIF signature is in line with our previously published data involving spontaneous or enhanced (via metformin treatment) resolution in young mice^15,21,34 as well as other recent studies^26,29,30,34. We also previously demonstrated downregulation of PLIN2, CEBPA, and PPARG in lung homogenates from IPF lung tissues compared with donors indicating reduced LIF abundance/differentiation²¹. In the current work, we expanded the LIF gene panel to other marker genes shown to constitute a signature for LIFs according to single-cell atlases, and the data showed significant downregulation at the lung homogenate level as well as downregulation of the LIF signature in the stromal fraction of various fibrotic lung diseases contained in the HLCA. Altogether, our aged fibrosis mouse model and human-derived data strongly indicate that LIFs are indeed impaired in pulmonary fibrosis.

Importantly, our preclinical model identified Adamts4 + AFs that are associated with delayed fibrosis resolution. While bioinformatic analysis predicted that Fgf10+ cells do not necessarily pass through a MyoFB intermediate before giving rise to Adamts4+ cells, this intriguing aspect remains an open question that requires further experimental testing. During homeostasis, the balance between ECM synthesis and degradation is crucial for maintaining proper lung physiology and architecture. However, such ECM homeostasis is disrupted during lung fibrosis. Metalloproteinases play pivotal roles in regulating ECM turnover and fibrosis progression. MMPs exhibit both pro-fibrotic (for e.g., MMP-1, MMP-7, and MMP-9) and anti-fibrotic (e.g., MMP-2 and MMP-3) activities. Elevated levels of MMP-1, MMP-7, and MMP-9 have been observed in the blood serum and lung tissue of IPF patients, with MMP-7 considered a potential diagnostic biomarker^46,47,48. Of note, ADAMTS4 has been implicated in lethal immunopathology following influenza virus injury³³. VCAN is a major substrate for ADAMTS4, and it is a component of the provisional matrix that facilitates recruitment of inflammatory cells such as leukocytes^33,49. TIMP-3 inhibits ADAMTS4 but also other metalloproteinases. Timp3 expression is significantly reduced during fibrosis formation in young mice but recovers during fibrosis resolution. Timp3 knockout mice exhibit persistent inflammation and fibrosis, resembling the pathology observed in aged mice. This effect is attributed to sustained MMP activity, increased neutrophil infiltration, and impaired anti-inflammatory responses. Datamining of spatial transcriptomic data revealed that TIMP3 is among the top downregulated genes in IPF compared with donor. Interestingly, MMP-7 activity is elevated in Timp3 knockout mice, while Mmp7 knockout mice demonstrate impaired neutrophil migration and attenuated fibrosis after bleomycin injury^50,51. Our data using rTIMP-3 or rVCAN in the context of alveolar organoids and PCLS revealed strong attenuation of fibrogenesis. As both candidates are not ADAMTS4-specific, we also employed a genetic approach and showed that siRNA-mediated knockdown of ADAMTS4 yields similar antifibrotic effects. Importantly, datamining of the HLCA revealed that ADAMTS4 is among the top upregulated genes in IPF, ILD, and NSIP compared with the healthy state. Spatial transcriptomic data also demonstrated the significant upregulation of ADAMTS4 in IPF, and its correlation with fibrogenesis. These data strongly suggest that the relentless course of fibrogenesis seen in human pulmonary fibrosis may be due to continued inflammation and remodeling mediated by ADAMTS4 and VCAN degradation. Our ex vivo investigations revealed that rTIMP-3 or rVCAN treatment attenuates fibrogenesis at least in part by downregulating Adamts4 at the RNA and protein levels. To our knowledge, this phenomenon has not been described in the literature so far. Adamts4 downregulation might be due to a feedback mechanism involving cues originating from the ECM being modulated by rTIMP3 or rVCAN. As expected, activity assays showed reduced ADAMTS4 activity upon rTIMP-3 or siRNA-mediated ADAMTS4 knockdown. It is also possible that excess rVCAN saturates the protease activity of ADAMTS4, thus attenuating its detrimental effect. Another possibility is that VCAN acts directly on mesenchymal cells similarly to what has been shown in the context of cardiac repair⁵². It might also be that applying excess amounts of intact VCAN dilutes the profibrotic effect of VCAN fragments generated by ADAMTS4-mediated degradation. Thorough analysis of ECM modulation by ADAMTS4, its regulators, and its substrates as well as the associated contribution to sustained lung damage is an important aspect for further research. It will also be important to screen for small molecule inhibitors to selectively inhibit ADAMTS4 in the context of fibrogenesis and resolution using murine and humanized models of pulmonary fibrosis.

To conclude, our lineage-tracing approach identified Fgf10 + AFs with LIF signature as a cellular source of not only MyoFBs that arise during fibrogenesis but also Adamts4+ cells that persist during delayed fibrosis resolution. Datamining and ex vivo therapeutic interventions indicated strong clinical relevance for our findings and highlighted ADAMTS4 as a promising therapeutic target in pulmonary fibrosis. Our study provides valuable insights into the cellular origins and trajectories of AFs and MyoFBs and sets the stage for identifying further targets for future interventions.

Methods

Animal experiments

All animal studies were conducted in accordance with the ARRIVE guidelines and in compliance with the approved protocols by the animal ethics committee of Justus-Liebig-University Giessen and the local authorities (Regierungspraesidium Giessen, permit numbers: G76/2020–No. 1026_GP and 825_M). All mice were kept under specific pathogen-free (SPF) conditions with unrestricted access to food and water. Housing was regulated at 22 ± 2 °C with 55 ± 10% relative humidity and a 14/10-h light/dark cycle. Eight-to-twelve-week-old wild-type (WT) mice with a C57BL/6 genetic background were obtained from Charles River (#632). Since aged female mice show better survival and less severe lung fibrosis in response to bleomycin instillation compared with aged male mice³², they were chosen for the experiments. To lineage-label Fgf10+ cells, 52-to-56-week-old female Fgf10^Cre-ERT2/+; tdTomato^flox mice were used. These mice were previously generated and validated in our lab⁴⁵.

Human-derived lung tissues

Fresh human lung tissues from non-IPF donors used for precision-cut lung slice (PCLS) cultures were obtained from Hannover Medical School (MHH, Hannover, Germany) in compliance with ‘The code of ethics of the world medical association’ (Approval by the Ethics Committee of MHH renewed on 2015/04/22; approval number 2701-2015) and those from IPF patients from the UGMLC Giessen Biobank, member of the DZL Platform Biobanking (Approval by the Ethics Committee of Justus-Liebig-University Giessen; approval number AZ 58/15). Formalin-fixed, paraffin-embedded human donor and IPF lung sections used for in situ hybridization and fresh material used for RNA extraction followed by qPCR were obtained from the UGMLC Giessen Biobank (Approval by the Ethics Committee of Justus-Liebig-University Giessen; approval number AZ 58/15). The use of human samples was in accordance with the Declaration of Helsinki and written informed consent was obtained from all participants or their next of kin.

Tamoxifen and bleomycin treatments

To label Fgf10+ cells, mice were intraperitoneally (i.p.) injected with tamoxifen (Sigma-Aldrich, T5648-5G) reconstituted in corn oil (Sigma-Aldrich, C8267) at a dose of 0.1 mg g⁻¹ body weight. Mice received three successive injections on days −13, −11 and −9 prior to intratracheal (i.t.) instillation of saline or bleomycin (1.5–2 U kg^-1 body weight; Bleomedac, PNZ-02411351). Lungs were collected on days 14, 30 and 60 post bleomycin injury for further analysis.

Hematoxylin and eosin staining and fibrosis area quantification

Mouse lung perfusion was done using Hank’s Balanced Salt Solution (HBSS; Thermo Fisher Scientific, 14175095). Once harvested, the lungs were fixed with 4% paraformaldehyde (PFA; Merck, 104005) and sequentially dehydrated through a series of ethanol concentrations. The lungs were then embedded in paraffin and cut into 5-µm-thick sections using a microtome (Leica, RM2255). Before staining, paraffin sections were deparaffinized with xylol and sequentially rehydrated. The slides were stained with hematoxylin (Roth, T865.2) for 2 min, followed by eosin (Thermo Fisher Scientific, 6766007) for 1 min. The stained slides were dehydrated and mounted. The damaged area was quantified by Orbit image analysis software (V3.64, Idorsia Pharmaceuticals Ltd).

Picro Sirius Red and Masson’s trichrome staining

Five-µm-thick paraffin sections were deparaffinized and rehydrated. For Picro Sirius Red staining (Abcam, ab150681), an adequate amount of dye was applied to cover the tissue and incubated for 1 h at room temperature (RT). The dye was washed in two changes of 1% acetic acid. Masson’s Trichrome staining (Abcam, ab150686) was carried out according to the manufacturer’s instructions. Stained sections were dehydrated, mounted and finally imaged.

RNA extraction and qPCR

RNA was extracted using Trizol (Thermo Fisher Scientific, 10296028) and RNeasy mini kit (Qiagen, 74106). Complementary DNA (cDNA) was synthesized from total RNA using the QuantiTect reverse transcription kit (Qiagen, 205311) according to the manufacturer’s instructions. Quantitative real-time PCR (qPCR) was performed using PowerUp SYBR green master mix (Thermo Fisher Scientific, A25742) and a QuantStudio™ 3 Real-Time PCR System machine (Applied Biosystems™). Primer sequences are listed in Supplementary Data 1.

ADAMTS4 activity assay

To isolate protein, human or murine PCLSs were lysed using 150 μL of RIPA buffer (50 mM Tris-HCl pH 7.5, 0.15 M NaCl, 1% NP40). Briefly, PCLSs were cut into small pieces using scissors and incubated in 4 °C for 30 min. Then, the lysates were centrifuged at 10,000 × g for 10 min at 4 °C. Next, the supernatant was isolated and used for protein measurement with DC Protein Assay Reagents Package (BioRad) according to the manufacture’s protocol. ADAMTS4 activity was determined using the ADAMTS4 Fluorogenic Assay Kit (BPS Bioscience, 82546) according to the manufacturer’s instructions. Measurements were acquired using an Infinite M200 plate reader (Tecan) equipped with i-control 2.0 software (Tecan). Activity was calculated as follows to yield fluorescence units/μg:

$$\frac{{{{\rm{Sample\; reading}}}}-{{{\rm{Blank\,reading}}}}}{{{{\rm{Protein\,concentration}}}}}*1000$$

Fluorescence-activated cell sorting

Mice were euthanized and lungs were perfused with 20 mL HBSS through the right ventricle. To isolate tdTom+ cells for single-cell RNA sequencing (scRNA-seq), lungs from Fgf10^Cre-ERT2/+; tdTomato^flox mice were chopped into small pieces and digested using 0.5% collagenase type IV (Gibco, 17104-019) in HBSS for 30 min on a rotator at 37 °C. After digestion, lung homogenates were passed through 70 µm and 40 μm cell strainers (Sarstedt, 83.3945070 and 83.3945.040, respectively) to obtain single-cell suspensions. After centrifugation, the pellet was resuspended in an antibody cocktail containing anti-mouse CD31 (AF488-conjugated, Biolegend, 102514, 1:100), CD45 (AF488-conjugated, Biolegend, 103121, 1:100) and CD326 (EpCAM, APC/Cy7-conjugated, Biolegend, 118218, 1:50) for 20 min on ice in the dark. After washing, cells were stained with SyTOX (1:1000, Invitrogen, S34862) to exclude dead cells and sort live cells for scRNA-seq.

For murine alveolosphere assays, freshly isolated resident mesenchymal cells (rMCs) identified as CD45^neg CD31^neg EpCAM^neg Sca-1^pos cells were co-cultured with freshly isolated alveolar epithelial cells (AECs) as previously described⁵³. In brief, lungs from young WT mice were inflated i.t. with 3 mL dispase (5 U/mL; BD Biosciences, 354235) and further digested by incubation with 2 mL dispase solution (5 U/mL) for 40 min at RT. The cell suspension was incubated with biotin rat anti-mouse CD45 (BD Biosciences, 553078), CD16/32 (BD Biosciences, 553143), and CD31 (BD Biosciences, 553371) for 30 min to deplete hematopoietic and endothelial cells, respectively, using magnetic separation. The cell pellet was incubated in the antibody cocktail (anti-CD31, anti-CD45 and anti- CD326) containing gamma globulin (sandoglobulin) to sort CD45^neg CD31^neg EpCAM^low AECs. Anti-mouse Ly-6A/E (Pacific Blue-conjugated, Biolegend, 108120, 1:50) antibodies were used to isolate rMCs. FACSAria III cell sorter (BD Biosciences) was used for sorting and measurements. At least 100,000 rMCs and 400,000 AECs from one WT lung could be sorted. Data were analysed using FlowJo software (FlowJo LLC, version 10.10.0).

Alveolosphere cultures

The alveolosphere assay was performed as previously described³⁵. In brief, 17,000 FACS-sorted rMCs and 5000 AECs from young WT mice were resuspended in 50 µL organoid medium containing minimum essential medium (MEM) Alpha (Gibco, 41061029), 10% FBS, 1% L-glutamine-penicillin-streptomycin (Sigma-Aldrich, G1146), 1% insulin/transferrin/selenium (ITS) (Gibco, 41400-045) and 0.0002% heparin (Stem cell Technologies, 07980). The cells were mixed at a ratio of 1:1 with cold growth factor-reduced phenol red-free Matrigel (Corning, 356231), transferred to a 24-well plate on a 0.4 µm cell culture insert (Millipore, PICM01250) and incubated for 15 min for Matrigel polymerization. Next, 350 µL of organoid medium was added to the lower chamber. Cultures were incubated under air-liquid conditions at a temperature of 37 °C with 5% CO₂ for three weeks. The culture medium was changed every other day. Alveolospheres were imaged using EVOS M7000 (Thermo Fisher Scientific, AMF7000). The formed WT alveolospheres were treated with 5 ng/mL rTGFβ1 for three days, followed by rTIMP-3 (300 ng/mL) treatment for another three days. The alveolospheres were used for qPCR and IF. Primer sequences are listed in Supplementary Data 1. RNA extraction was done using micro kit (Qiagen, 74004).

Whole-mount immunofluorescence of alveolospheres

Cultured alveolospheres were fixed with 4% PFA for 20 min at RT and washed with 1X PBS. Samples were blocked and permeabilized with 0.5% Triton-X-100 and 5% donkey serum in 1X PBS overnight at 4 °C. Primary anti-Pro-SFTPC and mouse/rat anti-Rage (R&D Systems, MAB1179, 1:200) antibodies were diluted in the blocking buffer and alveolospheres were incubated overnight at 4 °C. Samples were washed three times using washing buffer (2% donkey serum, 0.3% Triton-X-100 in 1X PBS) followed by incubation with the corresponding secondary antibodies. Nuclei were subsequently counterstained with DAPI. The samples were covered with cover slips and ProLong^TM Glass Antifade Mountant (Thermo Fisher Scientific, P36984).

In situ hybridization, immunofluorescence and quantification

Five-µm-thick paraffin sections from mice or human transplanted lungs were deparaffinized and rehydrated. RNAscope was performed according to the manufacturer’s instructions (ACD, Doc. No. 323100-USM) using mouse probes for Fgf10 (Mm-Fgf10, 446371-C1), Inmt (Mm-Inmt, 486371-C1), Robo2 (Mm-Robo2, 475961-C2), Cthrc1 (Mm-Cthrc1, 413341-C1) and Adamts4 (Mm-Adamts4, 537341-C2) or human probes for CTHRC1 (Hs-Cthrc1, 413331-C1), ADAMTS4 (Hs-ADAMTS4, 497161-C3), INMT (Hs-INMT, 459961-C1) and ROBO2 (Hs-ROBO2, 502341-C2). RNAscope for Tcf21 (Mm-Tcf21, 508661-C1) was followed by immunofluorescence (IF) for SFTPC (Seven Hills Bioreagents, WRAB-9337, 1:1000), ACTA2 (Sigma, F3777-.2 ML, 1:100), or ADAMTS4 (Thermo Fisher Scientific, PA1-1749A, 1:150). Anti-tdTomato antibody (Sicgen, AB8181-200, 1:250) was used to boost the tdTomato fluorescent signal that was diminished after antigen retrieval. For IF, the samples were blocked with 5% BSA serum (Jackson ImmunoResearch, 001-000-162) for 1 h at RT, followed by incubation with primary antibodies overnight at 4 °C. The next day, samples were washed three times with PBS, then incubated with secondary antibodies for 1 h at RT. Nuclei were counterstained using 4´,6-diamidino-2-phenylindole (DAPI; Thermo Fisher Scientific, D1306) and slides were mounted by Fluoromount-G (Southern Biotech, 0100-01). Fluorescent images were acquired using Mica Widefocal LiveCell System (Leica microsystem). Images were quantified using Fiji software⁵⁴.

Generation of human and mouse precision-cut lung slices (PCLS) and treatments

Fresh human lung tissues from IPF patients that underwent lung transplantation were inflated with 2% UltraPure low melting agarose (Invitrogen, 16520-100) and cut into 300 µm-thick slices using a vibratome. The resulting human precision-cut lung slices (hPCLS) were cultured in DMEM/F-12, GlutaMAX^TM medium (Thermo Fisher Scientific, 31331028) supplemented with 10% fetal bovin serum (FBS) (Gibco, 10270-106), 1% penicillin-streptomycin (Thermo Fisher Scientific, 15070063) and 0.5% amphotericin (Sigma, A2942). Cultures were treated with vehicle, 300 ng/mL recombinant protein TIMP-3 (rTIMP-3) (R&D Systems, 973-TM-010) or 500 ng/mL recombinant versican (rVCAN) (abbexa, abx069660) in medium containing 0.1% serum for three days. Mouse PCLS were generated as described⁵⁵ at a thickness of 200 µm and treated with vehicle or 5 ng/mL rTGFβ1 (R&D Systems, 240-B) for five days followed by treatment with vehicle, rTGFβ1 + Vehicle or rTGFβ1 + rTIMP-3 for three days. PCLS obtained from Bleo d14 Fgf10^Cre-ERT2; tdTomato^flox mice were treated with vehicle or 500 ng/mL rVCAN for three days. The samples were collected for RNA extraction and qPCR.

Donor non-IPF PCLS were prepared from healthy margins of surgically resected lungs of patients with lung cancer. Briefly, human lung lobes were cannulated with a flexible catheter and the selected lung segments were inflated with warm (37 °C) low melting agarose (1.5%) prepared in Dulbecco’s Modified Eagle’s Medium Nutrient Mixture F-12 Ham (DMEM-F-12) supplemented with 15 mM HEPES, 100 U/mL penicillin, and 100 μg/mL streptomycin (all from Invitrogen Life Technologies). After polymerization of the agarose solution on ice, tissue cores of a diameter of 8 mm were prepared using a sharp rotating metal tube. Subsequently, the cores were sliced into 250−300 μm-thin slices in DMEM using a Krumdieck tissue slicer (Alabama Research and Development). PCLS were washed three times for 30 min in DMEM-F-12 supplemented with 15 mM HEPES, 100 U/mL penicillin, and 100 μg/mL streptomycin and used for the experiments.

ADAMTS4 gene silencing in PCLS

PCLS generated from human donor lungs were placed in 12-well-plates in 800 μL DMEM F-12 medium supplemented with HEPES and penicillin/streptomycin. Afterward, PCLS were transfected with lipoplexes of siLentFect lipid reagent (Bio-Rad Laboratories) encapsulating 100 nmol pre-designed, commercially available siRNA against ADAMST4 (Thermo Fisher Scientific, AM16830) in 200 μL OPTI-MEM. To control for non-specific gene inhibition of the siRNA used in this study, a universal negative-control scrambled siRNA (si-Scr) sequence was employed (Thermo Fisher Scientific, 4390843). The tissue slices were incubated for 48–72 h at 37 °C and 5% CO₂ and afterward collected for fixation (followed by immunofluorescence and in situ hybridization), RNA extraction (for qPCR), or protein isolation (for activity assays or western blots). Western blots were performed according to standard procedures using antibodies against ADAMTS4 (Thermo Fisher Scientific, PA1-1749A; 1:1000), COL1 (Southern Biotech, 1310-1; 1:500), and Beta-actin (ACTB; Sigma, A1978; 1:5000). The latter was used as a loading control. Secondary anti-rabbit (Dako, P0217; 1:2000), anti-goat (Dako, P0449; 1:2000), and anti-mouse (Dako, P0447; 1:5000) antibodies were used, respectively. Blots were developed using SuperSignal West Dura Extended Duration Substrate (Thermo Fisher Scientific, 34075) and imaged using ChemiDoc Touch Imaging System and Image Lab 6.1 software (Bio-Rad).

Single-cell RNA sequencing

To explore transcriptional changes specific to mesenchymal cells in the aged Fgf10^Cre-ERT2; tdTomato^flox animals, scRNA-seq was conducted on tdTom+ cells extracted from the lungs of saline or bleomycin-challenged mice. Cells were flow-sorted, and libraries were prepared using the Chromium Next GEM Single-Cell 3ʹ Reagent Kit v3.1 from 10x Genomics. Sequencing data underwent processing through a previously established pipeline⁵⁶, which in brief involved excluding cells with fewer than 1,200 detected counts or more than 40,000 counts, as well as those with over 20% mitochondrial gene expression. Genes detected in fewer than 20 cells were filtered out to ensure high-quality read alignment and gene quantification. Following the pre-processing of individual samples, data integration was performed using the Harmony algorithm, with batch correction applied through BBKNN (Batch Balanced K-Nearest Neighbors) and highly variable genes selected using SeuratV3 flavor.

Differentially expressed genes (DEGs) were identified using the presto package in Seurat, with a minimum log-fold change threshold of 0.5 and a detection rate of at least 50%. To compare differences between clusters and samples, differential expression analysis was performed using the FindMarkers function, followed by filtering and visualization of significant genes through a volcano plot using tidyplots⁵⁷ or with the pheatmap package. RNA velocity analysis was conducted to map transcriptional dynamics and cell state predictions were explored using scVelo package, selecting 500 highly variable genes and computing first- and second-order moments with 20 principal components and 30 nearest neighbors. Latent cell velocity and inferred differentiation trajectories were further inferred using CellRank⁵⁸, employing standard settings and vignette, enabling probabilistic modeling of cell-fate transitions. Cluster activity along latent time was quantified by mapping cluster identities onto the trajectory and visualizing smoothed density curves, highlighting early, transitional, and late contributors to the dynamic process. To gain functional insights, cluster markers were mapped to ENTREZ gene identifiers and annotated using the org.Mm.eg.db database. Pathway enrichment analyses were conducted using Gene Ontology (GO) and Kyoto Encyclopaedia of Genes and Genomes (KEGG) databases to identify enriched biological processes and signaling pathways, with results visualized using ggplot2 and dplyr.

Stromal cells from healthy, NSIP, IPF, and ILD donor lungs were obtained from the Human Lung Cell Atlas (HLCA)⁵⁹. Differential expression analysis was conducted by comparing IPF, NSIP and ILD to healthy stromal cells. DEGs were calculated by comparing each disease category against ‘normal’ (healthy stromal cells) using the Wilcoxon rank-sum test. Ensembl IDs mapped to the gene symbols were incorporated into the final data frame, and genes with infinite avg_log2FC values were excluded. Key transcriptional changes associated with the AF1 signature (Source Data file) were analysed using the clusterProfiler and fgsea packages.

Spatial transcriptomic analyses were performed using Squidpy⁶⁰. Visium libraries were subsetted while retaining spatial coordinates, and gene expression matrices were normalized prior to non-negative matrix factorization (NMF) to identify spatially structured transcriptional programs. NMF factor assignments were stored in the observation table and visualized with Squidpy, and factor loadings were compared to assess correspondence between spatial programs. Spot-level gene detection frequencies were quantified across conditions, group differences evaluated with Mann–Whitney U tests, and Pearson correlations of gene-positive spots computed to examine co-expression within each timepoint. All Visium datasets, including Space Ranger outputs and full-resolution images, were obtained from BioStudies (S-BSST1410, human)³⁸.

Statistical analysis and figure preparation

All data are presented as mean ± standard error of the mean (SEM). For comparison between two groups, student’s t test (unpaired, two-tailed) was employed. One-way ANOVA was used for comparisons among three or more groups with a single variable and two-way ANOVA was used for comparisons with two variables. Statistical tests are indicated in the corresponding figure legends. The P < 0.05 was considered statistically significant. Quantitative data were assembled and analyzed using GraphPad Prism (GraphPad Software, version 10.4.1 (532)). The number of biological replicates (depicted as n) is indicated in the corresponding figure legends. Figures were prepared and assembled using Adobe Illustrator (Adobe).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary

Data availability

Cluster quantification data, DEGs, gene signatures used for GSEA, and scoring are provided in the Source Data file. Raw and processed data were deposited on Gene Expression Omnibus (GEO) and are publicly available under the accession number GSE295566. Source data are provided with this paper.

Code availability

The analysis code is available through: https://github.com/alikhadim/ADAMTS4-Stromal-Remodeling/.

References

Gross, T. J. & Hunninghake, G. W. Idiopathic pulmonary fibrosis. N. Engl. J. Med. 345, 517–525 (2001).
Article CAS PubMed Google Scholar
Maher, T. M. Interstitial lung disease: a review. JAMA 331, 1655–1665 (2024).
Article CAS PubMed Google Scholar
Thannickal, V. J., Toews, G. B., White, E. S., Lynch, J. P. & Martinez, F. J. Mechanisms of pulmonary fibrosis. Annu Rev. Med. 55, 395–417 (2004).
Article CAS PubMed Google Scholar
Selman, M., Buendía-Roldán, I. & Pardo, A. Aging and pulmonary fibrosis. Rev. Investig. Clin. 68, 75–83 (2016).
CAS Google Scholar
Taskar, V. S. & Coultas, D. B. Is idiopathic pulmonary fibrosis an environmental disease? Proc. Am. Thorac. Soc. 3, 293–298 (2006).
Article PubMed Google Scholar
García-Sancho, C. et al. Familial pulmonary fibrosis is the strongest risk factor for idiopathic pulmonary fibrosis. Respir. Med. 105, 1902–1907 (2011).
Article PubMed Google Scholar
Raghu, G., Weycker, D., Edelsberg, J., Bradford, W. Z. & Oster, G. Incidence and prevalence of idiopathic pulmonary fibrosis. Am. J. Respir. Crit. Care Med. 174, 810–816 (2006).
Article PubMed Google Scholar
Thannickal, V. J. Mechanistic links between aging and lung fibrosis. Biogerontology 14, 609–615 (2013).
Article CAS PubMed PubMed Central Google Scholar
Hecker, L. et al. NADPH Oxidase-4 mediates myofibroblast activation and fibrogenic responses to lung injury. Nat. Med. 15, 1077–1081 (2009).
Article CAS PubMed PubMed Central Google Scholar
Sato, N. et al. Metformin attenuates lung fibrosis development via NOX4 suppression. Respir. Res. 17, 107 (2016).
Article ADS PubMed PubMed Central Google Scholar
Rangarajan, S. et al. Metformin reverses established lung fibrosis in a bleomycin model. Nat. Med. 24, 1121–1127 (2018).
Article CAS PubMed PubMed Central Google Scholar
Gabbiani, G., Ryan, G. B. & Majne, G. Presence of modified fibroblasts in granulation tissue and their possible role in wound contraction. Experientia 27, 549–550 (1971).
Article CAS PubMed Google Scholar
Hinz, B. et al. Recent developments in myofibroblast biology: paradigms for connective tissue remodeling. Am. J. Pathol. 180, 1340–1355 (2012).
Article CAS PubMed PubMed Central Google Scholar
Hinz, B. et al. The myofibroblast: one function, multiple origins. Am. J. Pathol. 170, 1807–1816 (2007).
Article CAS PubMed PubMed Central Google Scholar
Kheirollahi, V. et al. Metformin induces lipogenic differentiation in myofibroblasts to reverse lung fibrosis. Nat. Commun. 10, 2987 (2019).
Article ADS PubMed PubMed Central Google Scholar
El Agha, E. & Thannickal, V.J. The lung mesenchyme in development, regeneration, and fibrosis. J. Clin. Investig. 133, e170498. (2023).
Schultz, C. J., Torres, E., Londos, C. & Torday, J. S. Role of adipocyte differentiation-related protein in surfactant phospholipid synthesis by type II cells. Am. J. Physiol. Lung Cell Mol. Physiol. 283, L288–L296 (2002).
Article CAS PubMed Google Scholar
Park, J. et al. The Tcf21 lineage constitutes the lung lipofibroblast population. Am. J. Physiol. Lung Cell. Mol. Physiol. 316, L872–L885 (2019).
Article CAS PubMed PubMed Central Google Scholar
Al Alam, D. et al. Evidence for the involvement of fibroblast growth factor 10 in lipofibroblast formation during embryonic lung development. Development 142, 4139–4150 (2015).
PubMed PubMed Central Google Scholar
El Agha, E. et al. Fgf10-positive cells represent a progenitor cell population during lung development and postnatally. Development 141, 296–306 (2014).
Article CAS PubMed Google Scholar
El Agha, E. et al. Two-way conversion between lipogenic and myogenic fibroblastic phenotypes marks the progression and resolution of lung fibrosis. Cell Stem Cell 20, 261–273.e3 (2017).
Article CAS PubMed Google Scholar
Taghizadeh, S. et al. Characterization in mice of the resident mesenchymal niche maintaining AT2 stem cell proliferation in homeostasis and disease. Stem Cells 39, 1382–1394 (2021).
Article CAS PubMed Google Scholar
Chanda, D. et al. Mesenchymal stromal cell aging impairs the self-organizing capacity of lung alveolar epithelial stem cells. Elife 10, e68049 (2021).
Article CAS PubMed PubMed Central Google Scholar
Liang, J. et al. Lipid deficiency contributes to impaired alveolar progenitor cell function in aging and idiopathic pulmonary fibrosis. Am. J. Respir. Cell Mol. Biol. 71, 242–253 (2024).
Article CAS PubMed PubMed Central Google Scholar
Fang, Y. et al. RUNX2 promotes fibrosis via an alveolar-to-pathological fibroblast transition. Nature https://doi.org/10.1038/s41586-024-08542-2 (2025).
Trempus, C. S. et al. Functional Pdgfra fibroblast heterogeneity in normal and fibrotic mouse lung. JCI Insight 8, e164380 (2023).
Article PubMed PubMed Central Google Scholar
Tsukui, T., Wolters, P. J. & Sheppard, D. Alveolar fibroblast lineage orchestrates lung inflammation and fibrosis. Nature 631, 627–634 (2024).
Article ADS CAS PubMed PubMed Central Google Scholar
Tsukui, T. et al. Collagen-producing lung cell atlas identifies multiple subsets with distinct localization and relevance to fibrosis. Nat. Commun. 11, 1920 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Mayr, C. H. et al. Sfrp1 inhibits lung fibroblast invasion during transition to injury induced myofibroblasts. Eur. Respir. J. https://doi.org/10.1183/13993003.01326-2023 (2024).
Konkimalla, A. et al. Transitional cell states sculpt tissue topology during lung regeneration. Cell Stem Cell 30, 1486–1502.e9 (2023).
Article CAS PubMed PubMed Central Google Scholar
Sueblinvong, V. et al. Predisposition for disrepair in the aged lung. Am. J. Med. Sci. 344, 41–51 (2012).
Article PubMed PubMed Central Google Scholar
Redente, E. F. et al. Age and sex dimorphisms contribute to the severity of bleomycin-induced lung injury and fibrosis. Am. J. Physiol. Lung Cell Mol. Physiol. 301, L510–L518 (2011).
Article CAS PubMed PubMed Central Google Scholar
Boyd, D. F. et al. Exuberant fibroblast activity compromises lung function via ADAMTS-4. Nature 587, 466–471 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Lingampally, A. et al. Evidence for a lipofibroblast-to-Cthrc1 + myofibroblast reversible switch during the development and resolution of lung fibrosis in young mice. Eur. Respir. J. 65, 2300482 (2025).
Article CAS PubMed PubMed Central Google Scholar
Zabihi, M. et al. An optimized protocol for the generation of alveolospheres from wild-type mice. Cells 13, 922 (2024).
Article CAS PubMed PubMed Central Google Scholar
Kashiwagi, M., Tortorella, M., Nagase, H. & Brew, K. TIMP-3 is a potent inhibitor of Aggrecanase 1 (ADAM-TS4) and Aggrecanase 2 (ADAM-TS5)*. J. Biol. Chem. 276, 12501–12504 (2001).
Article CAS PubMed Google Scholar
Wayne, G. J. et al. TIMP-3 inhibition of ADAMTS-4 (Aggrecanase-1) is modulated by interactions between aggrecan and the C-terminal domain of ADAMTS-4. J. Biol. Chem. 282, 20991–20998 (2007).
Article CAS PubMed Google Scholar
Franzén, L. et al. Mapping spatially resolved transcriptomes in human and mouse pulmonary fibrosis. Nat. Genet. 56, 1725–1736 (2024).
Article PubMed PubMed Central Google Scholar
Vaccaro, C. & Brody, J. S. Ultrastructure of developing alveoli. I. The role of the interstitial fibroblast. Anat. Rec. 192, 467–479 (1978).
Article CAS PubMed Google Scholar
Barkauskas, C. E. et al. Type 2 alveolar cells are stem cells in adult lung. J. Clin. Investig. 123, 3025–3036 (2013).
Article CAS PubMed PubMed Central Google Scholar
Brownfield, D. G. et al. Alveolar cell fate selection and lifelong maintenance of AT2 cells by FGF signaling. Nat. Commun. 13, 7137 (2022).
Article ADS CAS PubMed PubMed Central Google Scholar
Liberti, D. C. et al. Alveolar epithelial cell fate is maintained in a spatially restricted manner to promote lung regeneration after acute injury. Cell Rep. 35, 109092 (2021).
Article CAS PubMed PubMed Central Google Scholar
Tahedl, D., Wirkes, A., Tschanz, S. A., Ochs, M. & Mühlfeld, C. How common is the lipid body-containing interstitial cell in the mammalian lung? Am. J. Physiol. Lung Cell Mol. Physiol. 307, L386–L394 (2014).
Article CAS PubMed Google Scholar
Schipke, J. et al. Lipofibroblasts in structurally normal, fibrotic, and emphysematous human lungs. Am. J. Respir. Crit. Care Med. 204, 227–230 (2021).
Article PubMed Google Scholar
Chu, X. et al. Validation of a Novel Fgf10Cre–ERT2 Knock-in Mouse Line Targeting FGF10Pos Cells Postnatally. Front. Cell Dev. Biol. 9, 671841 (2021).
Chuliá-Peris, L. et al. Matrix metalloproteinases and their inhibitors in pulmonary fibrosis: EMMPRIN/CD147 comes into play. Int. J. Mol. Sci. 23, 6894 (2022).
Article PubMed PubMed Central Google Scholar
Pardo, A., Cabrera, S., Maldonado, M. & Selman, M. Role of matrix metalloproteinases in the pathogenesis of idiopathic pulmonary fibrosis. Respir. Res. 17, 23 (2016).
Article PubMed PubMed Central Google Scholar
Rosas, I. O. et al. MMP1 and MMP7 as potential peripheral blood biomarkers in idiopathic pulmonary fibrosis. PLoS Med. 5, e93 (2008).
Article PubMed PubMed Central Google Scholar
Wight, T. N. Provisional matrix: a role for versican and hyaluronan. Matrix Biol. 60–61, 38–56 (2017).
Article PubMed Google Scholar
Gill, S. E. et al. Tissue inhibitor of metalloproteinases 3 regulates resolution of inflammation following acute lung injury. Am. J. Pathol. 176, 64–73 (2010).
Article CAS PubMed Google Scholar
Zuo, F. et al. Gene expression analysis reveals matrilysin as a key regulator of pulmonary fibrosis in mice and humans. Proc. Natl. Acad. Sci.USA 99, 6292–6297 (2002).
Article ADS CAS PubMed PubMed Central Google Scholar
Feng, J. et al. Versican promotes cardiomyocyte proliferation and cardiac repair. Circulation 149, 1004–1015 (2024).
Article CAS PubMed Google Scholar
Vazquez-Armendariz, A. I. et al. Multilineage murine stem cells generate complex organoids to model distal lung development and disease. EMBO J. 39, e103476 (2020).
Article CAS PubMed PubMed Central Google Scholar
Schindelin, J. et al. Fiji: an open-source platform for biological-image analysis. Nat. Methods 9, 676–682 (2012).
Article CAS PubMed PubMed Central Google Scholar
Khadim, A. et al. Protocol for the generation and immunostaining of thick murine lung tissue slices. STAR Protoc. 6, 104034 (2025).
Article CAS PubMed PubMed Central Google Scholar
Khadim, A. et al. Myofibroblasts emerge during alveolar regeneration following influenza-virus-induced lung injury. Cell Rep. 44, 115248 (2025).
Article CAS PubMed Google Scholar
Engler, J. B. Tidyplots empowers life scientists with easy code-based data visualization. iMeta 4, e70018 (2025).
Weiler, P., Lange, M., Klein, M., Pe’er, D. & Theis, F. CellRank 2: unified fate mapping in multiview single-cell data. Nat. Methods 21, 1196–1205 (2024).
Article CAS PubMed PubMed Central Google Scholar
Sikkema, L. et al. An integrated cell atlas of the lung in health and disease. Nat. Med. 29, 1563–1577 (2023).
Article CAS PubMed PubMed Central Google Scholar
Palla, G. et al. Squidpy: a scalable framework for spatial omics analysis. Nat. Methods 19, 171–178 (2022).
Article CAS PubMed PubMed Central Google Scholar

Download references

Acknowledgements

The authors thank Martin R Stampfli and Giuseppina Fascellaro for their critical input. We also acknowledge David Dippel for his technical assistance with the human precision-cut lung slice cultures and siRNA experiments, Kerstin Goth and Hannah Hofmann for the animal husbandry and genotyping, and Thomas Sontag and Ingrid Henneke for their help with the animal protocols. This study was supported by funding from CSL through their Research Acceleration Initiative (https://www.csl.com/csl-rai). A.L. was supported by an Institute for Lung Health (ILH) startup grant and the Excellence Cluster Cardio-Pulmonary System/Cardio-Pulmonary Institute (EXC 2026, project number 390649896). S.H. was supported by German Research Foundation (DFG) grants SFB-TR84 (project number 114933180, project B2), SFB1021 (project number 197785619, project C05), and KFO309 (project number 284237345, projects P2, P8); the Excellence Cluster Cardio-Pulmonary System/Cardio-Pulmonary Institute (EXC 2026, project number 390649896); the German Center for Lung Research (DZL, project number 82DZL005B1), the Hessen State Ministry of Higher Education, Research and the Arts (HMWK, Landes-Offensive zur Entwicklung Wissenschaftlich-ökonomischer Exzellenz, LOEWE, Förderlinie 4a project ID III L7–519/05.00.002). E.E.A. was funded by the Institute for Lung Health (ILH), DFG (EL 931/4-2, EL 931/5-1, EL 931/4-1, EL 931/2-2 (KFO309 284237345), and SFB 1213 Project-ID 268555672), the Excellence Cluster Cardio-Pulmonary System/Cardio-Pulmonary Institute (EXC 2026, project number 390649896), and DZL.

Funding

Open Access funding enabled and organized by Projekt DEAL.

Author information

These authors contributed equally: Mahsa Zabihi, Ali Khadim.

Authors and Affiliations

Department of Medicine V, Internal Medicine, Infectious Diseases and Infection Control, Universities of Giessen and Marburg Lung Center (UGMLC), German Center for Lung Research (DZL), German Center for Infection Research (DZIF), Justus-Liebig-University Giessen (JLU), Giessen, Germany
Mahsa Zabihi, Ali Khadim, Arun Lingampally, Georgios-Dimitrios Panagiotidis, Jan Halweg, Susanne Herold & Elie El Agha
Cardio-Pulmonary Institute (CPI), Giessen, Germany
Mahsa Zabihi, Ali Khadim, Arun Lingampally, Stefan Hadzic, Georgios-Dimitrios Panagiotidis, Daniel Kalina, Jan Halweg, Tara Procida-Kowalski, Marek Bartkuhn, Janine Koepke, Christos Samakovlis, Norbert Weissmann, Andreas Günther, Werner Seeger, Susanne Herold, Malgorzata Wygrecka & Elie El Agha
Institute for Lung Health (ILH), Giessen, Germany
Mahsa Zabihi, Ali Khadim, Arun Lingampally, Stefan Hadzic, Georgios-Dimitrios Panagiotidis, Daniel Kalina, Jan Halweg, Tara Procida-Kowalski, Marek Bartkuhn, Janine Koepke, Christos Samakovlis, Norbert Weissmann, Andreas Günther, Werner Seeger, Susanne Herold, Malgorzata Wygrecka & Elie El Agha
Department of Medicine II, Internal Medicine, Pulmonary and Critical Care, Universities of Giessen and Marburg Lung Center (UGMLC), German Center for Lung Research (DZL), Justus-Liebig-University Giessen, Giessen, Germany
Arun Lingampally, Stefan Hadzic, Georgios-Dimitrios Panagiotidis, Daniel Kalina, Janine Koepke, Christos Samakovlis, Norbert Weissmann, Andreas Günther, Werner Seeger & Malgorzata Wygrecka
Transdisciplinary Research Area Life and Health, Organoid Biology, Life & Medical Sciences Institute, University of Bonn, Bonn, Germany
Ana Ivonne Vazquez-Armendariz
CSL Innovation GmbH, Marburg, Germany
Daniel Kalina, Mario Boehm & Malgorzata Wygrecka
Biomedical Informatics and Systems Medicine, Justus-Liebig-University Giessen (JLU), Giessen, Germany
Tara Procida-Kowalski & Marek Bartkuhn
Oujiang Laboratory (Zhejiang Lab for Regenerative Medicine, Vision and Brain Health), School of Pharmaceutical Science, Wenzhou Medical University, Wenzhou, Zhejiang, China
Xuran Chu & Elie El Agha
School of Pharmaceutical Sciences, Wenzhou Medical University, Wenzhou, Zhejiang, China
Xuran Chu
Institute for Pathology, Hannover Medical School, Hannover, Germany
Peter Braubach
Department of Pulmonary and Critical Care Medicine, The Quzhou Affiliated Hospital of Wenzhou Medical University, Quzhou People’s Hospital, Quzhou, Zhejiang, China
Saverio Bellusci
Laboratory of Extracellular Matrix and Regeneration, Justus-Liebig-University Giessen (JLU), Cardio-Pulmonary Institute (CPI), German Center for Lung Research (DZL), Institute for Lung Health (ILH), Giessen, Germany
Saverio Bellusci

Authors

Mahsa Zabihi
View author publications
Search author on:PubMed Google Scholar
Ali Khadim
View author publications
Search author on:PubMed Google Scholar
Arun Lingampally
View author publications
Search author on:PubMed Google Scholar
Ana Ivonne Vazquez-Armendariz
View author publications
Search author on:PubMed Google Scholar
Stefan Hadzic
View author publications
Search author on:PubMed Google Scholar
Georgios-Dimitrios Panagiotidis
View author publications
Search author on:PubMed Google Scholar
Daniel Kalina
View author publications
Search author on:PubMed Google Scholar
Jan Halweg
View author publications
Search author on:PubMed Google Scholar
Tara Procida-Kowalski
View author publications
Search author on:PubMed Google Scholar
Marek Bartkuhn
View author publications
Search author on:PubMed Google Scholar
Xuran Chu
View author publications
Search author on:PubMed Google Scholar
Janine Koepke
View author publications
Search author on:PubMed Google Scholar
Christos Samakovlis
View author publications
Search author on:PubMed Google Scholar
Mario Boehm
View author publications
Search author on:PubMed Google Scholar
Norbert Weissmann
View author publications
Search author on:PubMed Google Scholar
Andreas Günther
View author publications
Search author on:PubMed Google Scholar
Werner Seeger
View author publications
Search author on:PubMed Google Scholar
Peter Braubach
View author publications
Search author on:PubMed Google Scholar
Susanne Herold
View author publications
Search author on:PubMed Google Scholar
Malgorzata Wygrecka
View author publications
Search author on:PubMed Google Scholar
Saverio Bellusci
View author publications
Search author on:PubMed Google Scholar
Elie El Agha
View author publications
Search author on:PubMed Google Scholar

Contributions

Conceptualization, E.E.A.; Investigation, M.Z., A.K., A.L., M.Ba., S.B. and E.E.A.; Methodology, M.Z., A.K., A.L., A.I.V-A., St.H., G.D.P., D.K., J.H., T.P.K., M.Ba., X.C., J.K., N.W., M.W., S.B. and E.E.A.; Writing the animal protocol, A.L.; Resources, C.S., M.Bo., N.W., A.G., W.S., P.B., S.H., M.W., S.B. and E.E.A.; Writing—Original Draft, M.Z., A.K., A.L. and E.E.A.; Writing—Review & Editing, A.K. and E.E.A.; Funding Acquisition, E.E.A.; Supervision, S.B. and E.E.A.

Corresponding authors

Correspondence to Arun Lingampally, Saverio Bellusci or Elie El Agha.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Vera Krymskaya, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information (download PDF )

Description of Additional Supplementary Files (download PDF )

Supplementary Data 1 (download DOCX )

Reporting Summary (download PDF )

Transparent Peer Review file (download PDF )

Source data

Source Data (download XLSX )

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Zabihi, M., Khadim, A., Lingampally, A. et al. Persistence of alveolar fibroblast-derived ADAMTS4+ cells in a preclinical model of delayed pulmonary fibrosis resolution. Nat Commun 17, 4205 (2026). https://doi.org/10.1038/s41467-026-72419-3

Download citation

Received: 23 April 2025
Accepted: 27 March 2026
Published: 08 May 2026
Version of record: 08 May 2026
DOI: https://doi.org/10.1038/s41467-026-72419-3