BCL9 inhibition promotes fibroblast lipogenesis by regulating macrophage–fibroblast interactions to attenuate pulmonary fibrosis

Abstract

Idiopathic pulmonary fibrosis (IPF) is a progressive and fatal interstitial lung disease with an urgent need for novel therapeutic strategies. M2 macrophage-derived TGF-β1 promotes fibroblast myogenesis, contributing to IPF pathogenesis. Targeting macrophage polarization and fibroblast function thus represents an effective therapeutic approach for treating IPF. Here, we identify B-cell lymphoma 9 (BCL9) as a key upstream regulator implicated in IPF pathogenesis. We demonstrate that BCL9 drives the macrophage M2 program through the MerTK-ERK-SPP1 axis. Notably, pharmacological inhibition of BCL9 with our novel peptide, hsBCL9_Z96, effectively attenuates pulmonary fibrosis by reprogramming macrophage-fibroblast crosstalk. Specifically, BCL9 inhibition promotes fibroblast lipogenesis via TGF-β1 signaling, which in turn supports alveolar type 2 (AT2) cell expansion. This macrophage–orchestrated fibroblast phenotypic switch from myogenic to lipogenic is visually corroborated by spatial transcriptomic analyses and immunofluorescence staining of human lung tissues. Functionally, the pathological role of BCL9 and efficacy of hsBCL9_Z96 are validated in human cellular models, including IPF patient-derived cells, confirming its translational significance. Collectively, our findings not only elucidate a novel BCL9-driven macrophage–fibroblast–AT2 cell axis in IPF but also establish hsBCL9_Z96 as a promising first-in-class therapeutic candidate, providing a strong rationale for targeting BCL9-mediated Wnt signaling in clinical IPF treatment.

Introduction

Idiopathic pulmonary fibrosis (IPF) is a progressive interstitial lung disease that threatens public health and is characterized by interstitial fibrosis with spatially heterogeneous pathology.^1,2 Its etiology is unknown, and the disease is therefore classified as idiopathic. The average survival time after diagnosis is only 2–3 years, largely due to respiratory failure.³ The spatial and temporal heterogeneity of fibrosis poses a significant diagnostic and therapeutic challenge. Current antifibrotic therapies can only slow fibrosis progression but cannot reverse established pulmonary fibrosis.^4–6 This stark reality underscores the urgent, unmet need to explore the fundamental molecular drivers of IPF to identify novel therapeutic targets. Pulmonary fibrosis appears to be the result of the dysregulated interplay between immune cells and structural cells within the lung microenvironment. Among these, macrophages, owing to their plasticity and diversity, act not only as the first line of defense in the lung but also as key contributors to IPF pathogenesis.^7–10 They are generally classified into two phenotypes: classically activated (M1) and alternatively activated (M2) macrophages.¹¹ Macrophages, particularly M2 macrophage-derived transforming growth factor beta 1 (TGF-β1), have been shown to promote fibroblast-to-myofibroblast transdifferentiation (FMT) and excessive collagen deposition^8,12,13. Notably, macrophage depletion attenuates bleomycin (BLM)-induced IPF in a mouse model, whereas macrophage deficiency during the resolution phase exacerbates fibrosis,¹⁴ highlighting the phenotypic regulation of macrophages as a crucial therapeutic approach for treating IPF.

The Wnt/β-catenin signaling pathway is a highly conserved pathway fundamental to development, cell fate determination, and tissue homeostasis,¹⁵ and its pathological reactivation plays a critical role in regulating organ fibrosis, in part through macrophage phenotype reprogramming.^16–19 Since its dysregulation was first reported in 2003,²⁰ multiple nodes of the Wnt pathway have been implicated in driving fibrotic progression via distinct cellular mechanisms.^21–25 Decreased Wnt signaling reduces the M2 macrophage phenotype and protects against BLM-induced lung fibrosis in mice, supporting its therapeutic potential.^19,26,27 However, most Wnt inhibitors have been discontinued in preclinical or clinical studies owing to severe side effects associated with their mechanisms of action.^28–30 For example, targeting β-catenin itself or its binding to TCF/LEF often disrupts the pathway’s vital physiological roles in tissue stem cell maintenance and regeneration, leading to severe on-target toxicities.^28–30 This critical impediment highlights the necessity for a novel, more precise intervention strategy that can selectively inhibit the pathological arm of Wnt/β-catenin signaling in fibrosis while sparing its essential physiological functions.

A promising strategy to achieve this selectivity lies in targeting specific transcriptional co-activators within the Wnt signalosome. B-cell lymphoma 9 (BCL9), a key nuclear transcriptional coactivator of β-catenin, binds to the N-terminal arm of β-catenin via its homology domain 2—an interaction interface that does not overlap with those of other critical nuclear partners of β-catenin—and is crucial for Wnt/β-catenin signaling.³¹ This unique protein–protein interaction presents a compelling therapeutic opportunity: selectively disrupting the BCL9/β-catenin interaction could potently downregulate a subset of Wnt-driven pathogenic gene programs implicated in fibrosis, while potentially minimizing the catastrophic side effects associated with broad-pathway inhibition. Despite the established roles of BCL9 in Wnt/β-catenin-driven oncogenesis and development,^32,33 its specific function in the pathogenesis of IPF, especially regarding macrophage-mediated fibrosis and intercellular communication, remains unknown. Given this significant knowledge gap, we hypothesized that BCL9 serves as a critical mechanistic nexus linking pathogenic Wnt/β-catenin signaling to pro-fibrotic macrophage polarization in IPF, and that selectively targeting BCL9 represents a precise therapeutic strategy for fibrosis intervention.

Here, we demonstrated the upregulation of BCL9 in IPF and further identified its clinical relevance to disease pathogenesis. Genetic depletion or pharmacological inhibition of BCL9 using our novel peptide inhibitor, hsBCL9_Z96, inhibited M2 polarization via the Mer tyrosine kinase (MerTK)–extracellular signal-regulated kinase (ERK)–secreted phosphoprotein (SPP1) axis and protected mice from BLM-induced pulmonary fibrosis. Mechanistically, BCL9 inhibition enabled macrophages to orchestrate a myogenic-to-lipogenic phenotypic switch in fibroblasts via cell–cell interactions, a finding supported by reanalysis of spatial transcriptomic data and immunofluorescence (IF) staining for human IPF tissues, which facilitated AT2 cell expansion. Further validation using human cellular models, including IPF patient-derived cells, confirmed that BCL9 regulates fibroblast lipogenesis through macrophage crosstalk, highlighting the translational potential of hsBCL9_Z96. Collectively, our study suggests that targeting BCL9-driven Wnt signaling may offer a promising therapeutic strategy for IPF.

Results

Increased BCL9 levels correlated with IPF progression

We analyzed transcriptome data from patients with IPF (accession No.: GSE124685) to investigate whether aberrant Wnt signaling contributes to IPF progression. Wnt signaling activity and the expression of genes that positively regulate the Wnt pathway were upregulated in patients with IPF (Supplementary Fig. 1a, b). Consistently, lungs from fibrotic mice also showed a significant increase in β-catenin expression (Supplementary Fig. 1c). Given the role of BCL9 as a key transcriptional coactivator in the Wnt pathway, we next assessed its relevance to IPF. We reanalyzed scRNA-seq data from human samples.³⁴ Compared with other interstitial lung diseases, BCL9 expression showed a stronger association with pulmonary fibrosis (Fig. 1a), with this association being particularly prominent in macrophages relative to other myeloid cell types (Fig. 1b). Further interrogation of the GSE136831 dataset confirmed elevated BCL9 levels across macrophage subpopulations in IPF (Fig. 1c–f), with a slight increase in the BCL9-positive M2 macrophage subset (Fig. 1g). Importantly, upregulation of BCL9 in macrophages was a common feature in both clinical (IPF patients, Fig. 1h) and preclinical (fibrotic mice, Fig. 1i) settings.

Fig. 1.

Increased BCL9 levels correlated with IPF progression. Based on a publicly available scRNA-seq dataset: a BCL9 expression levels across various types of ILD. b BCL9 expression profile across myeloid cell subtypes in IPF patients. c–g Reanalysis of the GSE136831 dataset focused on IPF macrophages. To control for age as a confounder, only subjects aged ≥40 years were included in both the IPF and control groups. c Identification of macrophage subsets in IPF. d Density distribution of BCL9 expression within the macrophage population. e BCL9 expression levels across individual macrophages. f Average BCL9 expression levels in macrophages. g Proportion of BCL9-expressing macrophages. h Immunofluorescence staining for BCL9 (red) and CD68 (green) in human lung sections, Scale bar: 25 μm. i Immunofluorescence staining for BCL9 (red) and F4/80 (green) in mouse lung sections, Scale bar: 25 μm. P value by Wilcoxon rank-sum test (e, g). IPF idiopathic pulmonary fibrosis, ILD interstitial lung disease, COPD chronic obstructive pulmonary disease, scRNA-seq single cell RNA sequencing

Taken together, our results implicate elevated BCL9 expression levels in the pathogenesis of IPF, highlighting its clinical significance.

Loss of BCL9 in macrophages attenuates pulmonary fibrosis in mice

Given the elevated expression of BCL9 in profibrotic macrophages (Supplementary Fig. 1d), we proceeded to investigate its functional role in macrophages during IPF progression. We performed macrophage depletion–adoptive transfer assays in an established murine lung fibrosis model using wild-type (WT) mice, comparing the effects of transferring WT versus Bcl9-knockout (KO) macrophages (Supplementary Fig. 2a). Bcl9-KO macrophages were isolated from tamoxifen-treated Bcl9^fl/fl Cre-ERT2 mice (Supplementary Fig. 3a, b). Flow cytometry confirmed efficient macrophage depletion and verified that bone marrow-derived macrophages (BMDMs) from both WT and Bcl9-KO mice were successfully polarized toward the M2 phenotype prior to adoptive transfer (Supplementary Fig. 2b–d). Although macrophage depletion alleviated BLM-induced lung fibrosis, mice receiving WT BMDMs exhibited exacerbated fibrosis, as evidenced by lower survival rates, higher lung coefficients, more severe alveolar destruction, and increased collagen deposition compared to those in mice receiving Bcl9-KO BMDMs (Supplementary Fig. 2e–h).

To further delineate the macrophage-specific role of BCL9 in vivo, we generated macrophage-specific Bcl9 KO (Bcl9^fl/fl Lyz2^Cre) mice (Fig. 2a, and Supplementary Fig. 3a, c). Bcl9^fl/fl Lyz2^Cre mice exhibited reduced susceptibility to BLM induced lung fibrosis compared with WT mice, as evidenced by a significantly lower lung coefficient (Fig. 2b). The marked decrease in Ccn4, Ccn1, and Ccn2 expression indicated the inhibition of Wnt signaling in Bcl9^fl/fl Lyz2^Cre mice (Fig. 2c–e). Pathologically, macrophage-specific Bcl9 depletion alleviated pulmonary fibrosis, promoted alveolar structure restoration, and reduced collagen deposition (Fig. 2f–h). Moreover, the proportion of M2 macrophage infiltration was dramatically lower in the lung tissue of Bcl9^fl/fl Lyz2^Cre mice than that in WT mice (Fig. 2i), indicating that Bcl9 deficiency in macrophages suppresses M2 macrophage polarization. In addition, the stable proportion of M2 macrophages observed in Bcl9^fl/fl Lyz2^Cre mice further suggests that the protective effect is attributable to both a failure in the expansion of the M2 macrophage population and a loss of the pro-fibrotic capacity of macrophages (Fig. 2i).

Fig. 2.

Genetic depletion of Bcl9 in macrophages attenuates pulmonary fibrosis in mice. a Scheme for the generation of mice with macrophage-specific Bcl9 knockout (Bcl9^fl/fl Lyz2^Cre). b Lung coefficients of WT and Bcl9^fl/fl Lyz2^Cre mice (n = 3). qPCR was performed to detect c Ccn4, d Ccn1, and e Ccn2 expression in lung tissues (n = 3). f Representative whole-lung sections from mice of each group (n = 3). g Representative image of H&E and Masson staining in lungs (n = 3). Scale bar: 500 μm. h Quantification of the degree of fibrosis based on pathological sections (n = 3). i Flow cytometric analysis of M2 macrophage infiltration in the lungs (n = 3). Data are derived from three independent experiments, and are represented as the mean ± SD. * P < 0.05, *** P < 0.001 by two-way ANOVA. BLM bleomycin, IPF idiopathic pulmonary fibrosis, i.t. intratracheal administration, H&E hematoxylin-eosin, WT wild-type

Taken together, these results indicate that loss of BCL9 in macrophages attenuates lung fibrosis in mice.

BCL9 orchestrates FMT by regulating macrophage phenotype reprogramming

To investigate whether BCL9 regulates macrophage polarization, we used an in vitro model in which mouse peritoneal macrophages (PMs) were polarized toward M1 or M2 phenotypes using specific stimuli (Fig. 3a, and Supplementary Fig. 4). Loss of Bcl9 in PMs dramatically inhibited M2 polarization compared to WT PMs, whereas Bcl9 overexpression (OE) promoted the M2 phenotype (Fig. 3b–d). Given that macrophages are a major source of the profibrotic mediator TGF-β1 during IPF progression,³⁵ we next examined TGF-β1 levels in the supernatant of M2 macrophages following Bcl9 KO or OE. Consistent with the polarization results, Bcl9 loss markedly reduced TGF-β1 secretion, whereas Bcl9 OE substantially increased TGF-β1 levels (Fig. 3e). Since TGF-β1 is a key fibroblast activator in the fibrogenic program,¹³ we designed an in vitro coculture system containing macrophages and mouse lung fibroblasts (MLFs) to determine whether BCL9 affects FMT by regulating macrophage polarization (Fig. 3f). In cocultures with Bcl9-KO PMs, MLFs showed downregulated expression of profibrotic genes, including Acta2, Col1a1, and Col3a1 (Fig. 3g–i). In contrast, MLFs cocultured with Bcl9-OE PMs exhibited significant upregulation of these genes (Fig. 3j–l).

Fig. 3.

Macrophage-intrinsic BCL9 drives M2-like reprogramming and promotes fibroblast-to-myofibroblast transdifferentiation. a Schematic representation of the generation of WT, Bcl9-KO, Bcl9-OE and Vector-OE PMs. Macrophages were polarized toward the M1 phenotype with LPS and IFN-γ or toward the M2 phenotype with IL-4. b Gating strategy for M2 macrophages by flow cytometry. Flow cytometry quantification of M2 macrophage percentages in c WT versus Bcl9-KO and d Vector-OE versus Bcl9-OE PMs after stimulation (n = 3). e TGF-β1 levels in culture supernatants from the indicated M2 macrophage groups, measured by ELISA (n = 3). f Illustration of a coculture system containing PMs and MLFs. qPCR analysis of fibrotic gene expression, including g Acta2, h Col1a1, and i Col3a1, in MLFs cocultured with WT or Bcl9-KO PMs (n = 3). qPCR analysis of fibrotic gene expression, including j Acta2, k Col1a1, and l Col3a1, in MLFs cocultured with Vector-OE or Bcl9-OE PMs (n = 3). Data are derived from three independent experiments, and are represented as the mean ± SD. * P < 0.05, ** P < 0.01, *** P < 0.001 by two-way ANOVA (c, d), or Student’s t test, unpaired, two-tailed (e, g–l). WT wild-type, KO knockout, OE overexpression, PMs peritoneal macrophages, LPS lipopolysaccharide, IFN-γ Interferon-γ, IL-4 Interleukin-4, ELISA enzyme-linked immunosorbent assay, MLFs mouse lung fibroblasts

Collectively, these results indicate that macrophage-intrinsic BCL9 is a key regulator of phenotypic reprogramming. BCL9 deficiency suppresses M2 polarization, ultimately attenuating the TGF-β1-driven fibrogenic program in fibroblasts.

hsBCL9_Z96 attenuates pulmonary fibrosis in mice

Based on our findings that depletion of macrophage-intrinsic BCL9 reduces lung fibrosis, we employed hsBCL9_Z96, a stapled peptide that specifically disrupts the BCL9/β-catenin interaction, to mechanistically interrogate the role of BCL9 and evaluate its therapeutic potential in IPF. hsBCL9_Z96 is a potent and selective research tool that inhibits Wnt signaling without significant toxicity.³² Functional assessment using BrdU assays confirmed that hsBCL9_Z96 specifically inhibited proliferation in Wnt-dependent cell lines, and qPCR analysis further showed that it did not transcriptionally regulate key genes of other major signaling pathways (NF-κB, GPCR, EGFR, and Notch), supporting its on-target specificity (Supplementary Fig. 5a, b). To assess its pharmacological effects, we first investigated whether pharmacological inhibition of BCL9 regulates macrophage polarization and thereby affects FMT in vitro. IL-4-stimulated macrophages polarized toward the M2 phenotype showed elevated CD206 expression, which was inhibited by hsBCL9_Z96 in a dose-dependent manner (Supplementary Fig. 6a). Correspondingly, measurement of TGF-β1 secretion in macrophage supernatants revealed that 2.5 μM hsBCL9_Z96 slightly inhibited TGF-β1 release, while 5 μM hsBCL9_Z96 significantly suppressed its secretion (Supplementary Fig. 6b). In a macrophage-fibroblast coculture system, hsBCL9_Z96 inhibited FMT, as evidenced by a slight decrease in Acta2 and a marked reduction in Col1a1, Col3a1, and Fn1 (Supplementary Fig. 6c–f), confirming that hsBCL9_Z96 suppresses FMT by attenuating M2 macrophage polarization.

We next focused on whether hsBCL9_Z96 attenuates lung fibrosis in vivo (Fig. 4a). Consistent with its in vitro activity, hsBCL9_Z96 treatment protected against BLM-induced alveolar destruction and reduced collagen deposition (Fig. 4b, c), accompanied by a dramatic reduction in lung coefficients, Ashcroft scores, and Szapiel scores (Fig. 4d–f). Moreover, hsBCL9_Z96 downregulated the expression of both Wnt-related genes (Ccn4, Ccn1, and Ccn2) and fibrotic genes (Col1a1, Col3a1, Tgfb1, and Arg1), serving as evidence for its dual action on Wnt pathway inhibition and fibrosis attenuation (Fig. 4g–m). Importantly, hsBCL9_Z96 administration reduced M2 macrophage infiltration into lung tissues (Fig. 4n).

Fig. 4.

hsBCL9_Z96 administration attenuates pulmonary fibrosis in mice. a Experimental scheme of BLM-induced pulmonary fibrosis and therapeutic hsBCL9_Z96 intervention. b Representative whole-lung sections from each group of mice (n = 6–10). c Representative images of H&E and Masson staining in lungs (n = 6–10). Scale bar: 625 μm. d Lung coefficient of mice (n = 6–10). Quantification of the e degree of fibrosis and f degree of inflammation based on pathological sections (n = 6–10). qPCR detection was used to determine the expression of genes, including g Ccn4, h Ccn1, i Ccn2, j Col1a1, k Col3a1, l Tgfb1, and m Arg1, in lung tissues (n = 4). n Flow cytometric quantification of the percentage of M2 macrophages in lung tissues (n = 4). Data are derived from three independent experiments, and are represented as the mean ± SD. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001 by Student’s, unpaired, two-tailed t test. BLM bleomycin, i.t. intratracheal administration, i.p. intraperitoneal injection, H&E hematoxylin-eosin

Collectively, these results indicate that pharmacological inhibition of BCL9 represents a viable therapeutic strategy for IPF. hsBCL9_Z96, a novel inhibitor of the BCL9/β-catenin complex, effectively attenuates fibrosis both in vitro and in vivo.

hsBCL9_Z96 reprograms the macrophage phenotype during IPF

To investigate the underlying mechanism by which hsBCL9_Z96 attenuates IPF, we performed scRNA-seq on lung tissues (Fig. 5a). All cells were categorized into nine clusters after dimensionality reduction and clustering using t-distributed stochastic neighbor embedding (t-SNE) analysis (Fig. 5b). Consistent with its therapeutic effect, hsBCL9_Z96 treatment significantly downregulated the expression of fibrotic genes (Acta2, Col1a1, Col1a2, and Col3a1) and suppressed the activity of fibrosis-related pathways (Wnt signaling, Hippo signaling, and focal adhesion) (Fig. 5c, and Supplementary Fig. 7), thereby confirming its antifibrotic activity. We next focused on the impact of hsBCL9_Z96 on macrophage dynamics. After extracting macrophages from the nine clusters, we found that hsBCL9_Z96 not only reduced the proportion of M2 macrophages (Fig. 5d) but also suppressed their profibrotic phenotype, as evidenced by reduced expression of profibrotic factors including Ctnnb1, Arg1, Tgfb1, Spp1, and Mertk (Fig. 5e). To delineate the differentiation trajectory, we performed pseudotime trajectory analysis using the Monocle 2 algorithm, tracing macrophage development from monocytes during IPF. In fibrotic lungs, macrophages differentiated toward a profibrotic phenotype dominated by M2 macrophages, whereas hsBCL9_Z96 treatment altered this differentiation trajectory, reduced the proportion of M2 macrophages, and reprogrammed macrophages toward an M1 phenotype (Fig. 5f).

Fig. 5.

hsBCL9_Z96 alters macrophage differentiation trajectories during pulmonary fibrosis. a Schematic of the experimental workflow of scRNA-seq (n = 3). b t-SNE plot showing the distribution of nine cell clusters identified in the scRNA-seq data. c Dot plot showing the expression levels of fibrotic genes in the lungs. d Proportion of M2 macrophages within the total macrophage population. e Dot plot showing the expression levels of profibrotic genes in M2 macrophages. f Pseudotime analysis of the macrophage differentiation trajectories. Arrows indicate the inferred direction of differentiation. scRNA-seq, single-cell RNA sequencing

Taken together, these results indicate that pharmacological inhibition of BCL9 by hsBCL9_Z96 reprograms the macrophage phenotype and suppresses profibrotic macrophage activation during IPF.

Targeting BCL9 disrupts the MerTK–ERK–SPP1 axis to restrain profibrotic macrophage activation

To investigate the intrinsic gene expression patterns induced by hsBCL9_Z96, we extracted the gene expression profiles of M2 macrophages from the scRNA-seq data and visualized the top 10 genes using a heatmap (Fig. 6a). Notably, the most downregulated gene in the hsBCL9_Z96-treated group was Spp1 (red arrow in Fig. 6a), whose dynamic expression pattern, based on pseudotime trajectory analysis, paralleled the expression curve of Ctnnb1, indicating a regulatory relationship between Ctnnb1 and Spp1 following BCL9 inhibition by hsBCL9_Z96 (Fig. 6b). Given the established association between high SPP1 expression in macrophages and M2 polarization as well as lung fibrosis,^34,36 we speculated that BCL9 inhibition suppresses the profibrotic phenotype of macrophages by regulating SPP1 expression. Indeed, both pharmacological inhibition and genetic depletion of BCL9 suppressed SPP1 expression, as evidenced by reduced Spp1 expression in fibrotic lungs in vivo (Fig. 6c) and decreased SPP1 secretion from M2 macrophages in vitro (Fig. 6d, e). Importantly, silencing Spp1 using Spp1 siRNA eliminated the ability of hsBCL9_Z96 to inhibit M2 polarization and the secretion of the profibrotic factor TGF-β1 (Fig. 6f, g). The knockdown efficiency of Spp1 siRNA was confirmed via qPCR (Supplementary Fig. 8a). These data demonstrate that BCL9 inhibition regulates the macrophage phenotype through an SPP1-dependent mechanism.

Fig. 6.

BCL9 inhibition suppresses M2 macrophage polarization through the MerTK-ERK-SPP1 axis. a Heatmap of differentially expressed genes in M2 macrophages. b Correlation analysis of Ctnnb1 and Spp1 expression dynamics along the pseudotime trajectory. c qPCR was used to detect the Spp1 expression in the lungs (n = 4). ELISA measured SPP1 protein levels in culture supernatants from d PMs treated with or without hsBCL9_Z96 or e BMDMs from WT or Bcl9^fl/fl Cre-ERT2 mice (n = 3). f MFI of the M2 marker CD206 on PMs (n = 3). g TGF-β1 content in the supernatant of PMs was measured by ELISA (n = 3). Flow cytometric analysis of MerTK and p-ERK MFI in h PMs treated with or without hsBCL9_Z96 and i BMDMs from the indicated genotypes (n = 3). j Flow cytometric analysis of CD206 MFI in PMs from the indicated groups (n = 3). k SPP1 and TGF-β1 content in the supernatant of PMs (n = 3). Flow cytometric analysis of l CD206 MFI and m p-ERK MFI in PMs (n = 3). ELISA measurement of n SPP1 and o TGF-β1 levels in PMs supernatants (n = 3). Data are derived from three independent experiments, and are represented as the mean ± SD. ns, no significance. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001 by Student’s t test, unpaired, two-tailed (c, d, f–h, j–o), or two-way ANOVA (e, i). BLM bleomycin, ELISA enzyme-linked immunosorbent assay, PMs peritoneal macrophages, BMDMs bone marrow-derived macrophages, KO knockout, MFI mean fluorescence intensity, IL-4 interleukin-4, ERKi ERK inhibitor, p-ERK phosphorylated ERK

Next, we sought to better define the molecular mechanism by which BCL9 inhibition regulates macrophage polarization through SPP1. We hypothesized that MerTK-mediated phosphorylation of ERK may serve as a key signaling intermediate in this process, as macrophages coexpressing SPP1 and MerTK are markedly increased in IPF,³⁶ and MerTK expression and ERK activation are essential for SPP1 maturation and secretion.³⁷ Flow cytometry analysis demonstrated that BCL9 inhibition suppressed both MerTK expression and ERK phosphorylation (p-ERK) during M2 polarization (Fig. 6h, i). Notably, ERK inhibitor treatment during M2 polarization attenuated hsBCL9_Z96-mediated inhibition of the M2 phenotype and decreased SPP1 and TGF-β1 secretion (Fig. 6j, k), indicating that hsBCL9_Z96 regulates the macrophage phenotype through an SPP1-dependent, p-ERK-mediated mechanism. We then performed siRNA-mediated knockdown of MerTK in macrophages (Supplementary Fig. 8b) to elucidate the role of MerTK in BCL9 inhibition-mediated macrophage phenotype alteration. As expected, hsBCL9_Z96-induced suppression of CD206 expression, p-ERK levels, and SPP1 and TGF-β1 secretion was abolished upon MerTK siRNA treatment (Fig. 6l–o), indicating that MerTK regulation is an indispensable intermediate step in macrophage phenotype reprogramming driven by BCL9 inhibition.

Collectively, these data suggest that BCL9 inhibition, whether achieved through pharmacological inhibition or genetic depletion, regulates macrophage phenotype via the MerTK–ERK–SPP1 axis, leading to a decreased proportion of M2 macrophages and a reduced secretion of TGF-β1 during IPF.

BCL9 inhibition drives the myogenic–to–lipogenic phenotypic switch in fibroblasts through macrophage crosstalk during IPF

Given that BCL9 regulates FMT in fibroblasts across macrophages in vitro, validating this mechanism in vivo is a key focus of subsequent research. Fibroblasts extracted from the scRNA-seq data were categorized into four clusters (Supplementary Fig. 9a), including lipofibroblasts–lipid droplet-rich fibroblasts adjacent to AT2 cells, involved in alveolar maturation and surfactant production.³⁸ Since recent studies suggest that lipofibroblasts serve as precursor cells for myofibroblasts in BLM-induced pulmonary fibrosis,³⁹ with TGF-β1 promoting their transdifferentiation,⁴⁰ we hypothesized that the hsBCL9_Z96–driven reduction in TGF-β1 secretion from M2 macrophages could shift fibroblasts toward a lipogenic phenotype. Consistent with this hypothesis, CellPhoneDB analysis revealed that signaling between M2 macrophage-derived TGF-β1 and its receptors on fibroblasts was significantly suppressed after hsBCL9_Z96 treatment, particularly the TGFB1-TGFBR2 interaction (Fig. 7a). Pseudotime trajectory analysis further showed that fibroblasts predominantly underwent a lipogenic–to–myogenic transition in response to BLM stimulation, whereas hsBCL9_Z96 treatment reversed this trajectory, promoting a myogenic-to-lipogenic transition (Fig. 7b). Accordingly, hsBCL9_Z96 upregulated lipogenic genes in both myofibroblasts and lipofibroblasts and downregulated myogenic genes in myofibroblasts (Fig. 7c).

Fig. 7.

hsBCL9_Z96 promotes lipogenesis in fibroblasts through macrophage–fibroblast crosstalk. a Cell-cell interaction analysis between M2 macrophages and fibroblasts through the TGFB1 axis. b Pseudotime analysis of fibroblast differentiation trajectories. Arrows indicate the inferred direction of differentiation. c Dot plot showing the expression levels of lipogenic and myogenic genes in myofibroblasts and lipofibroblasts. d Reanalysis of a public spatial transcriptomic dataset from IPF patients. e Immunofluorescence staining of human lung tissues (n = 5). ADFP (red), CD68 (green), CD163 (purple), DAPI (blue), Scale bar: 25 μm. f Lipid droplet staining (LipidTOX, green; DAPI, blue) in MLFs cultured with conditioned media from PMs (n = 3). Scale bar, 130 μm, red arrow: lipid droplet. g Pparg expression in MLFs under the indicated co-culture conditions (n = 3). Flow cytometric analysis of the proportion of h lipofibroblasts and i AT2 cells in lung tissues from WT and Bcl9^fl/fl Lyz2^Cre mice (n = 3). Flow cytometric analysis of the proportion of j lipofibroblasts and k AT2 cells in lung tissues from mice treated with or without hsBCL9_Z96 (n = 3). l Proportion of AT2 cells within the total lung cells analyzed from scRNA-seq data. m Immunofluorescence staining of lung tissues from mice (n = 3). SPC (purple), DAPI (blue), Scale bar: 25 μm. Data are derived from three independent experiments (f–k, m). Data are represented as the mean ± SD. ns, no significance. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001 by Student’s t test, unpaired, two-tailed (g, j, k), or two-way ANOVA (h–i). scRNA-seq single cell RNA sequencing, MLFs mouse lung fibroblasts, PMs peritoneal macrophages, AT2 alveolar epithelial type 2 cells, WT wild-type

To spatially contextualize these interactions, we reanalyzed publicly available spatial transcriptomic datasets⁴¹ and performed IF staining on human lung tissues. This integrated analysis revealed aberrant accumulation of macrophages, lipofibroblasts, and myofibroblasts in IPF patients (Fig. 7d), with M2 macrophages found to be spatially adjacent to both lipofibroblasts and myofibroblasts within fibrotic regions (Fig. 7d, e). This spatial proximity strengthens the evidence for possible intercellular communication between M2 macrophages and lipofibroblasts in the fibrotic microenvironment. In vitro coculture experiments using murine cells further confirmed that conditioned media from hsBCL9_Z96-treated M2 macrophages increased lipid-droplet accumulation in MLFs (Fig. 7f) and dramatically upregulated Pparg expression (Fig. 7g). Notably, this effect was abolished following TGF-β1 neutralizing antibody treatment (Fig. 7g), suggesting that the pro-lipogenic influence of hsBCL9_Z96-treated macrophages depends on TGF-β1.

To assess the in vivo relevance, we quantified lipofibroblasts and AT2 cells, which depend on lipofibroblast support for function and niche maintenance,^38,42,43 in mice with macrophage-specific BCL9 depletion or after hsBCL9_Z96 treatment (gating strategy in Supplementary Fig. 9b). Both genetic and pharmacological BCL9 inhibition increased the proportion of lipofibroblasts and AT2 cells (Fig. 7h–k). The increase in AT2 cells was further confirmed by scRNA-seq data and IF staining, thereby controlling for the potential influence of airway epithelial cells (Fig. 7l, m, and Supplementary Fig. 9c).

Taken together, these data demonstrated that BCL9 inhibition promotes a myogenic–to–lipogenic phenotypic switch in fibroblasts through macrophage crosstalk, thereby increasing the proportion of AT2 cells, which collectively contribute to fibrosis resolution in IPF.

Patient-derived and human cell line models validate the translational significance of hsBCL9_Z96 in pulmonary fibrosis

To assess the translational significance of hsBCL9_Z96, we performed in vitro assays using human cellular models: macrophages derived from the human monocytic cell line THP-1 or IPF patient PBMCs, along with the human fetal lung fibroblast 1 (HFL1) cell line. Consistent with our murine data, hsBCL9_Z96 inhibited M2 polarization in THP-1 cells (Fig. 8a) and suppressed the MerTK-ERK-SPP1 axis (Fig. 8b–d). Moreover, hsBCL9_Z96-treated THP-1 cells promoted a myogenic-to-lipogenic phenotypic switch in HFL1 cells, as evidenced by decreased ACTA2 expression and increased lipid droplet accumulation (Fig. 8e, f). Through its regulation of macrophages, hsBCL9_Z96 also significantly regulated fibroblast proliferation and inhibited their invasive capacity (Supplementary Fig. 10).

Fig. 8.

hsBCL9_Z96 demonstrates translational potential by attenuating human macrophage-driven fibrosis. a Flow cytometric analysis of the proportion of M2 macrophages in THP-1-derived macrophages (n = 3). Representative immunofluorescence images showing b MerTK (red) and c p-ERK (green), in treated THP-1-derived macrophages; nuclei were counterstained with DAPI (blue) (n = 3), scale bar: 70 μm. d SPP1 levels in THP-1 supernatant (n = 3). e qPCR was used to detect ACTA2 expression in correspondingly treated HFL1 cells (n = 3). f Lipid droplet staining (LipidTOX, green; DAPI, blue) in HFL1 cells (n = 3). Scale bar, 70 μm, red arrow: lipid droplet. g Proportion of M2 macrophages in PBMC-derived macrophages from IPF patients (n = 5). qPCR detected the expression of h ACTA2, i COL1A1, j COL3A1, k PPARG, l SREBF2, and m PLIN2 in HFL1 cells cultured with conditioned media from the patient-derived macrophages described in g (n = 5). Data in a–f were derived from three independent experiments. Data are represented as the mean ± SD. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001 by one-way ANOVA (a, d, e) or Student’s t test, unpaired, two-tailed (g–m). PBMCs peripheral blood mononuclear cells, IPF idiopathic pulmonary fibrosis

Critically, in macrophages derived from IPF patient PBMCs, hsBCL9_Z96 treatment effectively reduced the M2 macrophage proportion (Fig. 8g). Furthermore, conditioned media from these hsBCL9_Z96-treated human macrophages profoundly attenuated HFL1 activation, as shown by reduced expression of profibrotic genes (ACTA2, COL1A1, and COL3A1) and induction of a lipogenic profile (increased PPARG, SREBF2, and PLIN2) (Fig. 8h–m).

Collectively, these results from human cellular models confirm that hsBCL9_Z96 inhibits M2 macrophage polarization via the MerTK-ERK-SPP1 axis and promotes fibroblast lipogenesis, thereby supporting its therapeutic potential in pulmonary fibrosis. Our findings provide compelling preclinical evidence that targeting BCL9 represents a promising strategy for treating IPF.

Discussion

The dual challenge of complex pathology and limited therapeutic options in IPF underscores the urgent need for novel molecular targets and the development of effective treatments. While alveolar epithelial injury remains central to IPF pathogenesis, M2 macrophage polarization contributes to exacerbated lung fibrosis by secreting profibrotic factors such as TGF-β1, which directly drives fibroblast myogenesis.^44–46 Concurrently, the Wnt/β-catenin signaling plays a crucial role in driving fibrosis progression by orchestrating distinct cellular functions—notably macrophage activation—which ultimately converge to promote fibrotic outcomes.^16,17,47,48 However, despite its therapeutic appeal, broad inhibition of Wnt signaling has been clinically problematic due to on-target toxicities associated with disrupting its physiological roles,²⁸ highlighting a critical need for more precise intervention strategies. This study addresses this unmet need by identifying BCL9, a transcriptional coactivator within the Wnt pathway, as a novel and targetable node in IPF. We demonstrated that macrophage-intrinsic BCL9 expression is elevated in human IPF and governs M2 polarization through the MerTK–ERK–SPP1 axis to attenuate pulmonary fibrosis. The suppression of the M2 phenotype following BCL9 inhibition reduced TGF-β1-driven myogenesis and promoted lipogenesis in fibroblasts through macrophage crosstalk, facilitating AT2 cell expansion (Fig. 9). We propose that targeting the BCL9/β-catenin interaction fulfills this need by offering a mechanism-based approach to inhibit a key fibrotic node while potentially avoiding the systemic side effects. Our findings that BCL9 drives M2 macrophage polarization align with and extend the established role of Wnt signaling in modulating immune cell function within the fibrotic niche. We mechanistically link BCL9 to a key pathogenic axis centered on SPP1^hi macrophages, a population whose abnormal aggregation of SPP1^hi macrophages in fibrotic lungs is strongly correlated with fibrosis progression^34,36 and is thought to activate myofibroblasts based on causal modeling.³⁶ Specifically, we demonstrate that BCL9 regulates SPP1 expression in an ERK-dependent manner downstream of MerTK, a pathway implicated in macrophage effector functions. These findings consistent with the colocalization of SPP1 and MerTK in macrophages and the established role of MerTK-mediated ERK phosphorylation in SPP1 secretion.^36,37 Our data positions BCL9 upstream of a critical pro-fibrotic secretory phenotype. Furthermore, the observed downregulation of TGF-β1 secretion upon BCL9 inhibition connects this Wnt coactivator to a master fibrogenic cytokine, providing a direct link to fibroblast activation. These results contextualize BCL9 within a known pathogenic cascade—Wnt signaling promotes M2 macrophages which secrete TGF-β1—and identify it as a specific regulatory component within this cascade. These mechanistic insights suggest BCL9 as a novel therapeutic target.

Fig. 9.

Schematic representation of the mechanism of BCL9 inhibition-mediated resolution of pulmonary fibrosis. BCL9 inhibition, achieved either by genetic knockout or via the peptide inhibitor hsBCL9_Z96, suppresses the pro-fibrotic phenotype of macrophages, thereby reprogramming macrophage–fibroblast crosstalk. This reprogramming drives a phenotypic switch in fibroblasts from a myogenic (fibrosis formation) to a lipogenic (fibrosis resolution) state. As a result, this transition not only attenuates pathological extracellular matrix (ECM) deposition but also promotes the expansion of alveolar epithelial type II (AT2) cells, collectively leading to the resolution of pulmonary fibrosis. Mechanistically, disruption of the BCL9/β-catenin interaction inhibits the MerTK–ERK–SPP1 axis in macrophages, leading to reduced secretion of TGF-β1. The ensuing attenuation of TGF-β1 signaling from macrophages to fibroblasts upregulates lipogenic genes (e.g., PPARG, PLIN2, SREBF2) while downregulating myofibroblast-associated genes (e.g., ACTA2, COL1A1, COL3A1) in fibroblasts. Together, these findings delineate a novel macrophage–fibroblast–AT2 cell communication axis through which targeting BCL9 orchestrates the resolution of lung fibrosis

A critical and novel insight from our study is the mechanism by which BCL9 inhibition in macrophages resolves fibrosis. Beyond simply reducing TGF-β1-driven myogenesis, reprogrammed macrophages actively instruct fibroblasts to undergo a phenotypic switch from a myogenic to a lipogenic state. Fibrosis formation involves a lipogenic–to–myogenic phenotypic shift in fibroblasts, whereas fibrosis resolution requires a myofibroblast–to–lipofibroblast transition, which is mediated by the suppression of TGF-β1 signaling and the activation of PPARγ signaling.³⁹ The regulation of BCL9 inhibition on fibroblast lipogenesis through macrophage–fibroblast crosstalk was supported by trajectory analysis, in vitro coculture, and crucially, by spatial transcriptomic and histological evidence from human IPF tissues showing a correlative niche for this crosstalk. The resultant lipofibroblast phenotype, characterized by the upregulation of PPARG, PLIN2, and SREBF2, is associated with reduced ECM deposition. Lipofibroblasts support alveolar maturation and surfactant production,³⁸ and PPARγ agonists increase the proportion of lipofibroblasts and AT2 cells to promote alveolar regeneration during lung injury.⁴⁹ Consistently, BCL9 inhibition-mediated lipogenesis increased the proportions of lipofibroblasts and AT2 cells in vivo during IPF. This alteration of AT2 cells may promote alveolar regeneration. Collectively, BCL9 inhibition does not merely halt a driver of fibrosis; it appears to actively promote a resolution-permissive microenvironment by reprogramming intercellular communication along a macrophage–fibroblast–AT2 cell axis.

The translational potential of targeting BCL9 is significant, primarily because the unique protein–protein interface of the BCL9/β-catenin complex presents a distinctive and druggable vulnerability. This interface does not overlap with binding sites for other essential nuclear co-factors,³¹ offering a rare opportunity for selective pharmacological disruption of pathogenic Wnt signaling without broadly inhibiting its physiological functions—a major limitation that has stalled the development of Wnt inhibitors. Our peptide inhibitor, hsBCL9_Z96, validates this therapeutic concept by demonstrating potent anti-fibrotic activity in vivo, and, importantly, the ability to induce fibroblast lipogenesis via macrophage crosstalk in human patient-derived cells. These results underscore its direct relevance to human disease and highlight a clear translational path forward. However, important questions remain for future research. First, whether Wnt pathway activation drives BCL9 upregulation or vice versa remains an important scientific question. Addressing this regulatory hierarchy will provide deeper insights into IPF pathogenesis. Second, while our data support a pro-regenerative role, the functional impact of increased lipofibroblasts and AT2 cells on alveolar regeneration requires formal validation, ideally using complex models like alveolar organoids. Finally, the long-term efficacy and safety profile of hsBCL9_Z96 warrant a comprehensive investigation in advanced preclinical models.

In summary, our study establishes a correlation between elevated BCL9 levels and the progression of IPF, highlighting its clinical relevance. We demonstrate that targeting BCL9 restrains M2 polarization through the MerTK–ERK–SPP1 axis to attenuate pulmonary fibrosis. Our novel peptide inhibitor, hsBCL9_Z96, emerges as a promising therapeutic candidate for IPF, promotes fibroblast lipogenesis through macrophage-fibroblast crosstalk, thereby facilitating AT2 cell expansion. Our study provides new mechanistic insights for developing innovative therapeutic strategies against IPF.

Materials and methods

Further description of all other methods is provided in the Supplementary Materials.

Collection of samples from patients with IPF

All human sample collections were conducted in accordance with ethical approval (K24-553) obtained from the Ethics Committee of Shanghai Pulmonary Hospital. Paraffin-embedded lung tissue sections from five patients diagnosed with IPF were used for immunofluorescence staining analysis. Normal lung tissue samples were obtained from patients undergoing pulmonary resection for solitary benign lesions and served as controls.

For the isolation of patient-derived peripheral blood mononuclear cells (PBMCs), peripheral blood was collected from five IPF patients. PBMCs were isolated using a human peripheral blood lymphocyte separation tube (7922112, Dakewe) according to the manufacturer’s instructions. The isolated PBMCs were further purified using CD14 MicroBeads (130-097-052, Miltenyi Biotec). These purified monocytes were then differentiated into mature macrophages by induction with 25 ng/mL recombinant human macrophage colony-stimulating factor (M-CSF) (574804, Biolegend) for 5 days and used for subsequent experiments.

Mice

Male C57BL/6J mice weighing about 25 g were purchased from Shanghai Slack Laboratory Animal Ltd, Co. Bcl9 ^fl/fl and Bcl9 ^fl/fl Cre-ERT2 mice were obtained from the laboratory of Prof. Basler in Switzerland. Lyz2^Cre mice were obtained from the laboratory of Prof. Huang Min in Shanghai Institute of Materia Medica, Chinese Academy of Sciences (SIMM). Macrophage-specific Bcl9 deficient mice (Bcl9^fl/fl Lyz2^Cre) were generated by intercrossing Bcl9 ^fl/fl with Lyz2Cre mice. Primers for mice genotyping are listed in Supplementary Table S1. All mice were housed in a specific pathogen-free facility of SIMM, with the relative humidity of 50 ± 10%, temperature of 25 ± 1 °C, a 12 h light/12 h dark cycle and diet and drinking water were provided ad libitum. For animal studies, adult (6–8 weeks) male mice were used and were randomly separated into groups. We ensured that experimental groups of animals were balanced in terms of age and weight. Animal care and experiments were approved by the Institutional Laboratory Animal Care and Use Committee (IACUC) of Shanghai Institute of Materia Medica, Chinese Academy of Sciences, with the approval number 2020-12-GLK-15.

Cells

Primary mouse peritoneal macrophages (PMs) were cultured in RPMI 1640 (Meilunbio, Dalian, China, MA0215) supplemented with 10% fetal bovine serum (FBS) (Thermo, Waltham, MA, USA, A5670701). Primary mouse bone marrow-derived macrophages (BMDMs) were cultured in fresh DMEM (Meilunbio, Dalian, China, MA0212) containing 20 ng/mL macrophage colony-stimulating factor (Biolegend, CA, USA, 576404) and 10% FBS (Thermo, Waltham, MA, USA, A5670701) for 5 days. To induce macrophage polarization, cells were treated with 100 ng/ml LPS (MERCK, Darmstadt, Germany L2630) + 20 ng/mL IFN-γ (GenScript, Nanjing, China, Z02916) to polarized towards M1 phenotype, and 40 ng/mL IL-4 (GenScript, Nanjing, China, Z02996) to polarized towards M2 phenotype. Primary mouse lung fibroblasts (MLFs) were cultured in DMEM (Meilunbio, Dalian, China, MA0212) containing 10% FBS (Thermo, Waltham, MA, USA, A5670701), 0.1 mg/mL streptomycin and 100 μ/ml penicillin (Meilunbio, Dalian, China, MA0110-1). Human fetal lung fibroblast cells (HFL1; IMMOCELL, Fujian, China, IM-H117) were cultured in F-12K medium (Thermo, Waltham, MA, USA) supplemented with 10% FBS. Human acute monocytic leukemia cells (THP-1; IMMOCELL, Fujian, China, IM-H260) were cultured in RPMI 1640 (Meilunbio, Dalian, China, MA0215) supplemented with 10% FBS. Colo 320DM human colorectal adenocarcinoma cells, MDA-MB-231 human breast adenocarcinoma cells, B16-F10 murine melanoma cells, and RKO human colon carcinoma cells were obtained from the American Type Culture Collection (ATCC) and maintained according to the supplier’s recommendations in standard culture media supplemented with 10% FBS and antibiotics. Cells were seeded 24 h before treatment and subsequently exposed to the indicated concentrations of hsBCL9_Z96, followed by incubation at 37 °C in a humidified atmosphere containing 5% CO₂ for an additional 24 h. All cells were incubated at 37 °C under 5% CO₂ in a humidified atmosphere.

Animal models and therapeutic experiments

Bleomycin (BLM) was used to induce a model of pulmonary fibrosis in mice as previously described,⁵⁰ but with some modifications. Specifically, mice were anesthetized with Avertin (Meilunbio, Dalian, China, MA0478) via intraperitoneal injection and received BLM (Selleck, Houston, TX, USA, S1214) via intratracheal injection to establish a model of pulmonary fibrosis. Control mice were treated with sterile saline by intratracheal injection. For therapeutic experiments, administration of hsBCL9_Z96, liposome-hsBCL9_Z96 and nintedanib were beginning after BLM injury according to an experimental schedule, in which hsBCL9_Z96 was administrated every 2 days by intraperitoneal injection, liposome-hsBCL9_Z96 was delivered every 2 days via intranasal administration, and nintedanib was administrated once a day by intragastric injection. The survival and death of mice were recorded every day during the experiment. At the end point of the experiment, mice were sacrificed, and the body weight and the lung tissue weight of the mice were recorded for analysis of the lung coefficient. After that, the inflated left lung tissues were fixed with buffered 10% formalin solution, and the right lung tissue was analyzed by qPCR. A flow cytometry assay was performed using the whole lung tissues.

Immunofluorescence staining

Human and murine lung tissue sections and cell slides were subjected to immunofluorescence staining using a Tyamide Signal Amplification (TSA) multiplex immunofluorescence kit (Huilanbio Biological Technology, Shanghai, China, RC0086plus-34RM) according to the manufacturer’s instructions. The antibody used are as follows: BCL9 (Proteintech, Wuhan, China, 22947-1-AP), CD68 (Abcam, Cambridge, MA, ab213363), F4/80 (Abcam, Cambridge, MA, ab300421), ADFP (Proteintech, Wuhan, China, 80362-2-RR), CD163 (Abcam, Cambridge, MA, ab182422), SPC (Abcam, Cambridge, MA, ab90716), MERTK (bcam, Cambridge, MA, ab182422), SPC (Abcam, Cambridge, MA, ab300136), p-ERK (CST, USA, 4370S).

Histological analysis

Lung tissues were fixed with buffered 10% formalin solution for 24 h at room temperature and then embedded in paraffin. Paraffin-embedded fixed lung tissues were cut into 4-μm slices for H&E and Masson staining. H&E staining kit (Solarbio, Beijing, China, G1120) and Masson’s trichrome staining kit (Solarbio, Beijing, China, G1340) were used for histological analysis. Hematoxylin–eosin (H&E) and Masson’s trichrome staining were used to quantify the degree of pulmonary fibrosis and inflammation based on Ashcroft and Szapiel scores.^51,52

Preparation of cell suspensions for single-cell RNA sequencing and flow cytometry

Single-cell suspensions were obtained from lung tissues after digested by collagenase I (Yeasen, Shanghai, China, 40507ES60), and filtered through 75 μm nylon mesh (Dakewe, Shanghai, China, 7061011). Red blood cell lysis buffer (Yeasen, Shanghai, China, 40401ES60) was then used to remove erythrocytes from single-cell suspensions.

Single-cell RNA-seq data analysis

Single-cell suspensions from lung tissues were loaded into the 10×Genomics Chromium Single Cell chips. Each barcoded cell sample’s RNA was reverse-transcribed and generating libraries which were prepared using the Chromium Single Cell 5′ GEM Library according to the manufacturer’s instructions. Final sequencing was performed with an Illumina Hiseq3000 according to the manufacturer’s instructions. All sequencing data were processed with Seurat (version 4.0.3). For quality-control step, cells with more than 20% mitochondrial features or with less than 1000 unique molecular identifier (UMI) were removed. Raw UMI counts were normalized with a scale factor of 10,000 UMI per cell and subsequently natural log transformed. For the remaining cells, the similarity and variability of cells were summarized by principal component analysis (PCA). The dimensionality of each dataset was further reduced using either the Barnes-Hut t-Distributed Stochastic Neighbor embedding (t-SNE). After that, cell types were identified using various known markers; cells expressing two or more specific markers were considered to be mixtures, which were removed before subsequent analysis. The 9 principal components were visualized and reanalyzed in order to identify subclusters. The function FindMakers was used to identify differentially expressed genes. Ligand–receptor interaction analysis of scRNA seq data was performed using the cellphoneDB function (Python 3.7) and ktplots R package. Pseudotime-trajectory was used to analyze the differentiation direction of target cells. Pseudotime-trajectory based on the Monocle 2 algorithm was used to understand the transcriptional dynamics of cells during pulmonary fibrosis.

Flow cytometry analysis

Single-cell suspensions were blocked with 4% FBS and anti-CD16/CD32 (BD Biosciences, San Jose, CA, USA, 553141) for 30 min at 4 °C. Cells were then incubated with surface marker antibodies for 20 min at 4 °C. Before intracellular labeling, the antibodies staining permeabilized with BD Cytofix/Cytoperm buffer (BD Biosciences, San Jose, CA, USA, 554714) was used according to the manufacture’s instructions. ACEA NovoCyte system was used to perform flow cytometry. Data analysis was done through NovoExpress software (version 1.4.0). The antibodies used are listed in Supplementary Table S2.

Quantitative real-time PCR

Total RNA from tissues or cell samples was isolated using Trizol reagent (Abclonal, Wuhan, China, RK30129) according to the manufacturer’s instructions. cDNA was reversed from RNA (1 μg) using ABScript III RT Master Mix (Abclonal, Wuhan, China, RK20428). SYBR Green Fast qPCR Mix (Abclonal, Wuhan, China, RK21203) was used to perform quantitative real-time PCR (qPCR). The relative quantity of target mRNA was normalized by use of GAPDH as an endogenous control, and calculated by the 2-(ΔΔCt) method. Primer sequences used are listed in Supplementary Table S3.

Coculture system of macrophages and lung fibroblasts

MLFs were seeded in the lower chamber of 24-well plates and incubated at 37 °C for 24 h to allow MLFs to adhere to the plates. PMs or the same volume of medium were added in the upper chamber of the transwell insert. After 24 h, PMs were induced to polarize towards the M1 or M2 phenotype upon treatment with or without hsBCL9_Z96. TGF-beta1,2,3 antibody (R&D, MN, USA, MAB1835) was used for neutralization assay according to the manufacturer’s instructions. After 48 h of coculture, MLFs in the lower chamber were collected for qPCR detection or lipid droplet staining.

Staining for lipid-droplet accumulation

Staining of lipid droplets in MLF was performed using HCS LipidTOX™ Green Neutral Lipid Stain (Invitrogen, CA, USA, H34475), nuclei stained with DAPI (Beyotime, Shanghai, China, C1002) according to the manufacturer’s instructions.

Enzyme-linked immunosorbent assay

The content of TGF-β1 and SPP1 in cell supernatant was measured by enzyme-linked immunosorbent assay (ELISA). ELISA was performed according to the manufacturer’s instructions. ELISA kits used are as follows: Mouse TGF-β1 ELISA Kit (Solarbio, Beijing, China, SEKM-0035), Mouse SPP1 ELISA Kit (MULTI SCIENCES, Zhejiang, China, EK2135).

Statistical analysis

All graphs and statistical analysis were performed using GraphPad Prism software (version 8.0.0; GraphPad Software, Inc., CA, USA). Data are presented as the mean ± SD. Two-tailed Student’s t tests were used for comparisons between two groups. Multiple groups were compared using one-way analysis of variance (ANOVA) or two-way ANOVA. The specific statistical method used for each experiment is indicated in the corresponding figure legends. P < 0.05 was considered as statistically significant. Asterisk coding is indicated in the figure legends as ^*P < 0.05, ^**P < 0.01, ^***P < 0.001.

Author contributions

Wenjie Wang: Conceptualization, resources, methodology, investigation, software, validation, formal analysis, data curation, visualization, writing-original draft, and writing-review, & editing. Yuan Zhang: Resources, methodology, investigation, funding acquisition and writing-review & editing. Fenglian He: Resources, methodology, investigation, software, validation, formal analysis and writing-review & editing. Li Sun: Resources, methodology, investigation, formal analysis, data curation, visualization, and writing-review & editing. Huiyu Li: Resources, investigation, formal analysis, data curation, visualization, and writing-review & editing. Rongchen Liu: Resources, investigation, formal analysis, data curation, visualization. Guanglin Zhong: Investigation, software, validation. Ling Zhang: Investigation. Anqi Li: Investigation, software, validation. Mei Feng: Methodology, investigation, visualization. Yuxuan Dong: Methodology, investigation, visualization. Xiaoxuan Lu: Methodology, investigation, visualization. Xiaojin Wang: Investigation, formal analysis, data curation, visualization. Yuan Si: Investigation, formal analysis, data curation, visualization. Yejun Wu: Investigation. Mengmeng Zhao: Investigation. Dafu Zhu: Investigation. Zhuyi Xi: Investigation, data curation, and visualization. Jian Chen: Investigation. Jing Chen: Conceptualization, supervision, project administration, and writing-review & editing. Ming-Wei Wang: Conceptualization, supervision, project administration, and writing-review & editing. Di Zhu: Conceptualization, supervision, project administration, and writing-review & editing. Likun Gong: Conceptualization, data curation, formal analysis, supervision, funding acquisition, project administration, and writing-review & editing. Bingshun Wang: Conceptualization, supervision, funding acquisition, project administration, and writing-review & editing. All authors have read and approved the article.

Data availability

scRNA-seq data is available at: 10.5281/zenodo.19181626. The complete dataset for this study has been uploaded and is accessible at: https://figshare.com/s/fab160d8f0d2c4cc5557.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41392-026-02753-x.