A novel protective role for SULF1 in chronic obstructive pulmonary disease through systemic compensation and binding to KRT15

Abstract

This study explores the role of Sulfatase 1 (SULF1) in the progression of Chronic Obstructive Pulmonary Disease (COPD). SULF1, a modifying enzyme involved in various biological processes, was quantified in human serum using ELISA, and its correlation with lung function was analyzed. The in vivo function of SULF1 was investigated through CRISPR-mediated gene knockout (KO) in mice, and SULF1 overexpression was induced via adeno-associated virus (AAV). A COPD model was established by exposing mice to cigarette smoke (CS). Our findings revealed that SULF1 levels were elevated in the serum of COPD patients, where they negatively correlated with the FEV₁/FVC ratio and the FEV₁/FVC ratio annual decline. In CS-exposed mice, SULF1 expression was downregulated in lung tissues but increased in serum and bronchoalveolar lavage fluid (BALF). SULF1 KO mice exhibited more pronounced emphysematous changes following CS exposure, whereas SULF1 overexpression partially mitigated these lung alterations. Additionally, SULF1 was found to play a protective role in COPD progression through stable binding with Keratin 15 (KRT15). In conclusion, this study elucidates the role of SULF1 in COPD progression, providing novel insights into its pathogenesis and identifying potential therapeutic targets.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12931-026-03818-9.

Introduction

Chronic Obstructive Pulmonary Disease (COPD) is a progressive airway disorder characterized by persistent airflow limitation [1, 2]. It currently ranks as the third leading cause of death globally, contributing significantly to the disease burden [3, 4]. In addition to smoking, genetic factors are key risk factors for COPD [5–7]. The pathological mechanisms underlying COPD are complex and not fully understood, involving various interconnected processes such as chronic inflammation, protease-antiprotease imbalance, and emphysema [8, 9]. These intricate processes involve multiple types of lung cells. Early COPD research primarily focused on airway epithelial cells [10, 11], but recent studies have increasingly recognized the critical role of lung fibroblasts, particularly in airway remodeling and inflammatory responses [12–15]. However, there remains a lack of core biomarkers that can specifically characterize pulmonary pathological changes and track disease progression in COPD, highlighting the need for further exploration in this field.

The Sulfatase 1 (SULF1) gene, the first identified member of the sulfatase family, encodes a heparan sulfate endosulfatase that selectively removes 6-O-sulfate groups from heparan sulfate (HS) chains in heparan sulfate proteoglycans (HSPGs) [16]. By altering the sulfation status of HSPGs, SULF1 regulates various physiological and pathological processes, including cell proliferation and tissue repair [17–20]. Most research on SULF1 has focused on its dual role in cancer, either promoting or suppressing tumor growth [21–25]. Studies on SULF1 in pulmonary diseases are limited, primarily focusing on lung cancer [26–28]. Recently, its involvement in pulmonary fibrosis has garnered attention, with evidence suggesting that SULF1 is upregulated in lung fibroblasts, driving fibrotic progression by catalyzing 6-O-desulfation of HS chains and activating TGF-β1 [29]. Conversely, downregulation of SULF1 in silica-induced human bronchial epithelial cells promotes proliferation and fibrotic responses [30]. However, research on the role of SULF1 in COPD remains scarce. Although Chen et al. identified SULF1 as a hub gene associated with COPD through bioinformatics analysis [31], there is a lack of systematic experimental studies on its role in COPD.

Keratin 15 (KRT15), a member of the type I keratin family, is predominantly expressed in the basal layer of epithelial tissues. It is an essential component of the cytoskeletal network, contributing to tissue integrity [32, 33]. KRT15 knockout (KRT15-/-) mice exhibit significantly exacerbated emphysema after long-term cigarette smoke exposure, indicating its protective role in maintaining alveolar structure [34]. These findings suggest that KRT15 may play a critical role in the pathogenesis and progression of COPD.

In this study, we identified SULF1 as a fibroblast-derived core gene associated with lung function in COPD by integrating single-cell RNA sequencing data from human and murine lung tissues with population-based transcriptomic datasets. We further performed in vivo functional studies using SULF1 knockout mice and an adeno-associated virus (AAV)–mediated SULF1 overexpression model under long-term cigarette smoke (CS) exposure, and confirmed that SULF1 exerts a protective role during COPD progression. Our findings showed that SULF1 expression was decreased in lung tissues but significantly increased in serum and bronchoalveolar lavage fluid (BALF) following CS exposure in mice. Consistently, serum SULF1 levels were markedly elevated in COPD patients and correlated with multiple pulmonary function indicators. Moreover, proteomic analysis revealed KRT15 as one of the differentially expressed proteins associated with SULF1, and their specific interaction within lung tissue was further validated. Collectively, these results suggest that SULF1 may participate in the regulation of emphysema progression through a compensatory peripheral circulating mechanism and interaction with KRT15, providing a potential therapeutic target for COPD precision medicine.

Results

Lung fibroblast-derived SULF1 is a core gene associated with lung function in COPD

Human (GSE173896: 3 healthy controls and 6 COPD patients) and murine (GSE168299: 4 air-exposed and 4 cigarette-exposed mice) lung tissue single-cell RNA sequencing datasets were retrieved from the GEO database (Fig. 1A). Annotation using established cell-specific marker genes identified 28 and 32 distinct cell subpopulations in the human and mouse datasets, respectively. Analysis revealed altered proportions of specific lung fibroblast subpopulations, including alveolar and adventitial fibroblasts, in both COPD patients and smoke-exposed mice (Fig. 1B). Further analysis demonstrated a consistent decrease in the prevalence of these two fibroblast subpopulations in both human COPD and murine smoke-exposure models (Fig. 1C), suggesting a potential protective role for lung fibroblasts in the pathogenesis of COPD.

Fig. 1.

Identification of 19 core COPD-associated regulatory genes in lung fibroblasts by integrated single-cell and transcriptomic analysis with machine learning. A Schematic workflow of scRNA-seq analysis. Human (GSE173896: 3 healthy controls and 6 COPD patients) and mouse (GSE168299: 4 air-exposed and 4 cigarette smoke (CS)-exposed mice) lung tissue scRNA-seq datasets were processed and analyzed. B) UMAP visualization showing annotated cell clusters in human and mouse lung tissues based on established marker genes, with alveolar and adventitial fibroblast subsets highlighted. C Proportional comparison of the two fibroblast subpopulations between COPD/CS groups and respective controls. D Heatmaps illustrating functional enrichment of consistently upregulated differentially expressed genes (DEGs) in fibroblast subsets. Each row represents a functional term or pathway, and each column represents an enriched gene set. Heatmap color scale indicates enrichment significance (−log10 adjusted p-value), with partially overlapping upregulated pathways highlighted. E Workflow for identifying COPD-associated regulatory genes combining transcriptome data (GSE47460: 91 non-COPD / 145 COPD; GSE76925: 40 non-COPD / 111 COPD) with weighted gene co-expression network analysis (WGCNA) using FEV1%pred as the phenotypic trait, followed by LASSO regression and experimental validation. F Four-way Venn diagramsillustrating cross-comparison between datasets to obtain overlapping genes from human or mouse adventitial fibroblasts (left) and alveolar fibroblasts (right). Genes appearing in at least three datasets were selected, resulting in 99 candidate genes. G Ten-fold cross-validation plots from LASSO regression analysis, with coefficient trajectories (left) and cross-validation error curves (right). The minimum lambda (λ) values for GSE47460-LASSO and GSE76925-LASSO were 51 and 43, respectively. H Venn diagram identifying 19 overlapping core regulatory genes shared betweenGSE47460-LASSO and GSE76925-LASSO models. I List of the 19 core regulatory genes

Differential gene expression analysis of the single-cell transcriptomic data (threshold: |log₂FC| > 0.5, p < 0.05) followed by gene set enrichment analysis revealed the upregulation of several biological pathways in lung fibroblasts from both human and mouse systems under COPD conditions (compared to respective controls). Notably, significantly enriched pathways included: R-HSA-1,280,215: Cytokine Signaling in Immune System, GO:0051336: Regulation of Hydrolase Activity, R-HSA-1,474,244: Extracellular Matrix Organization, GO:0045785: Positive Regulation of Cell Adhesion, GO:0048729: Tissue Morphogenesis, GO:0008283: Positive Regulation of Cell Population Proliferation, and GO:0001666: Response to Hypoxia (Fig. 1D).

We next analyzed human lung tissue transcriptomic datasets (GSE47460: 91 Non-COPD/145 COPD cases; GSE76925: 40 Non-COPD/111 COPD cases) and screened for differentially expressed genes (|log₂FC| > 0.5, p < 0.05). Weighted Gene Co-expression Network Analysis (WGCNA) was performed using FEV₁%pre as the lung function parameter of interest. Key modules containing differentially expressed genes were identified: the red, grey60, and lightgreen modules in GSE47460, and the black, pink, royalblue, and lightgreen modules in GSE76925 (Fig. 1E). Integration of the human lung tissue transcriptomic data (GSE47460 and GSE76925) with consistently significant differentially expressed genes from the two lung fibroblast subpopulations identified in the single-cell sequencing data yielded 99 feature genes(Fig. 1F). These genes were then analyzed using Lasso regression and generalized linear model (GLM) analysis to identify key variables. The GSE47460-LASSO and GSE76925-LASSO algorithms resulted in minimum lambda values of 51 and 43, respectively (Fig. 1G). The intersection of genes selected by both LASSO models identified 19 overlapping genes (Fig. 1H): AEBP1, BNC2, BST2, C3, COPZ2, ELN, GSN, HIVEP1, IGFBP6, PDGFD, PDGFRB, PLAT, RSPO3, SERPINE2, SPON1, SULF1, SVEP1, TMEM176B, and TXNDC5 (Fig. 1I).

Analysis of the GSE47460 and GSE76925 datasets revealed significant differential expression (p < 0.05) of most candidate genes between COPD and healthy control groups in human lung tissues (Fig. 2A). Furthermore, these genes were highly expressed in both human and mouse alveolar fibroblasts and adventitial fibroblasts (Fig. 2B). RT-qPCR validation confirmed that most candidate genes exhibited significant changes in mRNA expression levels in lung tissues from COPD patients and COPD mouse models compared to their respective controls (Fig. 2C). The primer sequences are provided in Table S1. In summary, by integrating the analysis of candidate gene expression from two independent human lung tissue transcriptomic datasets (GSE47460 and GSE76925) and RT-qPCR results from human and mouse lung tissues, we observed that SULF1 exhibited particularly remarkable changes, showing a 7.28-fold increase in COPD compared to non-COPD samples (Fig. 2D). This suggests that SULF1 may play an important role in the pathogenesis of COPD

Fig. 2.

Differential expression profiling and experimental validation of core COPD genes. A Bubble plot indicating expression patterns of candidate genes across diverse human and murine lung-resident cell types. Color scale denotes mean expression level, and bubble size reflects cell proportion expressing the gene. B RT-qPCR validation of gene expression in lung tissues from COPD patients and COPD mouse models (Mean ± SD; *p < 0.05, **p < 0.01, ***p < 0.001). C mRNA expression levels of candidate genes in lung tissues from COPD patients and controls in GSE47460 and GSE76925 datasets (Mean ± SD; ***p < 0.001, **p < 0.01, *p < 0.05). D Heatmap showing consistent expression patterns of the 19 core genes in mouse and human tissues comparing COPD and control samples. Rows represent genes, and columns represent sample groups (log2 fold change scale)

Elevated serum SULF1 levels in COPD patients are associated with lower baseline FEV1/FVC but slower longitudinal lung function decline

We next evaluated serum SULF1 levels in COPD patients and analyzed their associations with lung function. The clinical characteristics of the study population are shown in Table S2. A total of 410 subjects were included (COPD/Control = 186/224), with a mean age above 60 years. Regarding smoking status, 39.20% were current smokers, 20.24% were former smokers, and 40.00% were never-smokers. The smoking index was 17.83 ± 29.61 in the control group and 33.49 ± 32.72 in the COPD group. Serum SULF1 concentrations were significantly higher in COPD patients than in controls (p = 0.007) (Fig. 3A). This elevation remained consistent across different disease severity stages (GOLD 1, GOLD 2, GOLD 3/4) (p < 0.05) (Fig. 3B).

Fig. 3.

Serum SULF1 is elevated in COPD patients and correlates with clinical lung function metrics. A Comparison of serum SULF1 between control (n = 224) and COPD (n = 186) subjects (Welch’s t-test, p = 0.007). B Comparison of serum SULF1 among GOLD1, GOLD2, and GOLD3/4 subgroups (*p < 0.05, unpaired t-test). (C–H) Pearson correlation analysis between serum SULF1 and major pulmonary function indicators, showing correlation coefficients (r values) and p-values. Abbreviations: GOLD = Global Initiative for Chronic Obstructive Lung Disease; BD = bronchodilator; FVC = forced vital capacity; FEV₁ = forced expiratory volume in 1 s; %pred = percent predicted; LAA-950 = low-attenuation areas < − 950 HU (inspiration); LAA-856 = low-attenuation areas < − 856 HU (expiration); PRMfSAD = parametric response mapping for functional small airway disease

Regarding baseline lung function parameters, serum SULF1 levels were negatively correlated with the FEV₁/FVC ratio (Fig. 3C). In terms of longitudinal decline, serum SULF1 levels were significantly negatively correlated with the annual decline in FEV1/FVC (Fig. 3F). Negative correlations were additionally observed for the annual decline in FEV₁ and Post-bronchodilator (post-BD) FVC, although these did not reach statistical significance (Fig. 3D, E). No significant correlation was observed for the annual decline in FEV1% predicted. (Fig. 3G). In evaluating imaging-based air trapping parameters, serum SULF1 levels were positively correlated with expiratory LAA-856, whereas the correlations with inspiratory LAA-950 and PRMfSAD showed non-significant positive trends (Fig. 3H). Corresponding correlation coefficients and p-values are summarized in Table S3.

The results of the univariate linear regression analysis were largely consistent with the aforementioned findings (Table S4). In the multivariable linear regression model, after adjustment for potential confounding factors including age, gender, BMI and smoke status, higher serum SULF1 levels were independently associated with lower baseline pre- and post-BD FEV₁/FVC ratios (both P < 0.001),as well as with a slower annual decline in pre- and post-BD FEV₁ and FEV₁/FVC ratio (Table S5).

SULF1 protein is downregulated in lung tissues but upregulated in BALF and serum of COPD mouse models

We first assessed the expression of SULF1 in a COPD mouse model established through 6 months of cigarette smoke (CS) exposure. This exposure induced structural damage to the lung tissue and development of respiratory inflammation. H&E staining revealed that, compared to the control group, CS-exposed mice exhibited significant enlargement of alveolar septa and even alveolar fusion, accompanied by infiltration of inflammatory cells such as macrophages, and an increase in the mean linear intercept (MLI) (Fig. 4A). Compared with the AIR group, the CS group exhibited significantly higher destructive index, pathological score, airway wall thickness, and airway wall area (Figure S1A). In addition, CS exposure also promoted the accumulation of inflammatory cells, particularly neutrophils (Fig. 4B), and further induced increased secretion of the inflammatory cytokines IL-6 and CCL2 in the BALF (Fig. 4C). These findings demonstrate that prolonged CS exposure successfully induced pulmonary emphysema and airway inflammation in mice, confirming the establishment of a valid COPD model. RT-qPCR results showed that SULF1 mRNA expression was upregulated in the lung tissues of CS-exposed mice (Fig. 4D). In contrast, Western blot analysis revealed a significant reduction in SULF1 protein expression in the lung tissues of these mice (Fig. 4E). Moreover, we observed that SULF1 levels were significantly elevated in both BALF (Fig. 4F) and serum (Fig. 4G) of CS-exposed mice.

Fig. 4.

Evaluation of SULF1 expression levels in a COPD mouse model. A H&E staining of lung sections from CS-exposed and air-exposed mice for 6 months. Scale bar = 100 μm. Mean linear intercept (MLI) was used to quantify alveolar enlargement (n = 6). B Proportions of macrophages and neutrophils (percentage of total BALF cells) in AIR and CS groups (n = 3), Blue arrows denote neutrophils, whereas red arrows denote macrophages. C ELISA-based measurement of BALF IL-6, CCL2 (n = 6). D SULF1 mRNA expression levels in lung tissues of AIR and CS groups (n = 6). E Western blot analysis of SULF1 protein levels in lung tissues from AIR and CS groups (n = 6). F, G ELISA measurement of SULF1 levels in BALF and serum from AIR and CS groups (n = 6). Data are shown as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, and ns indicates not statistically significant. Statistical analysis was performed using Student’s t-test

CRISPR- mediated SULF1 knockout significantly aggravates COPD-like pathological phenotypes in mice

To further investigate the biological role of SULF1, we generated SULF1 knockout (KO) mice using CRISPR-Cas9 to delete a genomic segment, including exon 7 of the SULF1 gene (Figure S1B, C). Successful construction of the KO model was validated via DNA gel electrophoresis (Figure S1D), qRT-PCR (Fig. 5A), and Western blot analyses (Fig. 5B). Mice were then exposed to cigarette smoke (CS) or clean air for 6 months to establish COPD and control models (Fig. 5C).

Fig. 5.

SULF1 knockout(KO) exacerbated CS-induced COPD-like changes in mice. A, B Validation of SULF1-KO efficiency by qRT-PCR (A) and Western blot (B). C Protocol for stablishing COPD model using 6-month CS exposure. D Body weight comparisons among AIR-WT, AIR-KO, CS-WT, and CS-KO groups (n = 8–10). E H&E staining images comparing AIR-WT, AIR-KO, CS-WT, and CS-KO lung tissues (scale bar = 100 μm), with quantified MLI. F Pulmonary function measurements across groups (n = 5–6; (G) Immunohistochemical detection of α-SMA in lung sections from the four mouse groups and (H) quantitative comparison of α-SMA staining intensity (n = 6). Data are shown as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001,, and ns indicates not statistically significant. Statistical analysis was performed using Student’s t-test for panels (A) and (B), and one-way ANOVA followed by Bonferroni post-hoc testing for panels (D), (E), (F) and (H). Abbreviations: FRC, functional residual capacity; RI, resistance index; Cchord, chord compliance; FEV50/FVC, forced expiratory volume at 50% of vital capacity to forced vital capacity ratio; Cydn, dynamic compliance; MMEF, maximal mid-expiratory flow

Compared to the CS-exposed wild-type (CS-WT) group, SULF1 knockout mice (CS-KO) exhibited more pronounced weight loss (Fig. 5D). H&E staining showed that the CS-WT group displayed significant lung structural damage compared to the AIR-WT group, whereas these changes were further exacerbated in CS-KO mice (Fig. 5E). Lung function assessment (Fig. 5F) showed that CS-WT mice had significantly impaired lung function compared with AIR-WT mice, including increased functional residual capacity (FRC) and chord compliance (Cchord), and decreased FEV₅₀/FVC, dynamic compliance (Cdyn), and maximal mid-expiratory flow (MMEF). While respiratory impedance (RI) showed no significant differences. Compared with CS-WT mice, CS-KO mice exhibited further deterioration in lung function, characterized by reduced FEV₅₀/FVC and Cdyn, and increased Cchord (p < 0.05). No significant differences were observed in MMEF or RI between these two groups. Immunohistochemical staining of α-SMA indicated greater peribronchial α-SMA deposition in CS-KO mice than in CS-WT mice (Fig. 5G, H).These findings indicate that SULF1 deficiency exacerbates CS-induced lung injury and dysfunction, supporting a protective role of SULF1 in COPD progression.

AAV-mediated SULF1 overexpression significantly ameliorates partial pathological injuries and lung function decline in COPD model mice

To further investigate the functional role of SULF1 in COPD, we generated SULF1 overexpression plasmids (pAAV-SULF1) and control plasmids (pAAV-NC) (Fig. 6A). The efficacy of these plasmids was confirmed by transfecting 293T cells (Fig. 6B). Western blot analysis confirmed that pAAV-SULF1 significantly upregulated SULF1 protein expression (Fig. 6C). After 3 months of CS exposure, mice were administered either SULF1-overexpressing adeno-associated virus (AAV-SULF1) or control AAV (AAV-NC) via intratracheal instillation, followed by an additional 3 months of CS exposure (Fig. 6D). Western blot analysis demonstrated a marked increase in SULF1 protein expression in the lung tissues of AAV-SULF1 treated mice, confirming successful overexpression (Fig. 6E).

Fig. 6.

SULF1 overexpression alleviates CS-induced COPD-like changes in mice. A Construction of pAAV-SULF1 and pAAV-NC plasmids. B, C Validation of overexpression efficiency by plasmid transfection into 293T cells (B) and Western blot for FLAG-tagged SULF1 protein (C). D Timeline of CS exposure and AAV administration. E Western blot validation of SULF1 expression in lung tissues of CS-AAV-NC and CS-AAV-SULF1 mice. F Representative H&E staining of AIR-AAV-NC, CS-AAV-NC, and CS-AAV-SULF1 lung sections (scale bar = 100 μm), and MLI quantification (n = 6). G Lung function comparison among groups (n = 3–4). Mean ± SD; *p < 0.05, **p < 0.01, ***p < 0.001; ns = not significant; one-way ANOVA

H&E staining revealed that CS-exposed mice treated with AAV-NC exhibited more severe lung damage, including greater alveolar enlargement and increased MLI, compared to the AIR-AAV-NC group. In contrast, the CS-AAV-SULF1 group showed a significant reduction in alveolar enlargement and a lower MLI compared to the CS-AAV-NC group, although the MLI in the CS-AAV-SULF1 group remained significantly higher than that in the AIR-AAV-NC group (Fig. 6F). Lung function measurements also revealed that CS-AAV-NC mice exhibited significant impairment in lung function, characterized by increased FRC and Cchord, and decreased FEV₅₀/FVC and Cdyn, compared to the AIR-AAV-NC group. In contrast, the CS-AAV-SULF1 group showed significant improvement in lung function, with reduced FRC and Cchord, and increased FEV₅₀/FVC and Cdyn, compared to the CS-AAV-NC group. However, when compared to the AIR-AAV-NC group, the CS-AAV-SULF1 group still showed significantly elevated FRC, and no significant differences were observed in FEV₅₀/FVC, Cchord, or Cdyn (Fig. 6G). Immunohistochemical staining of α-SMA indicated greater peribronchial α-SMA deposition in CS-AAV-NC group than in AIR-AAV-NC mice, whereas this CS-induced α-SMA deposition was reduced in AAV-SULF1 group (Figure S3C).These findings suggest that SULF1 overexpression partially alleviated the structural damage induced and lung function decline in the COPD mouse model, though it does not fully reverse the pathological changes.

Spatial co-localization of SULF1 and KRT15 in the airway epithelial region and their interaction in lung tissue.

We performed RNA fluorescence in situ hybridization (RNA-FISH) combined with immunofluorescence (IF) experiments. These experiments revealed that SULF1 protein is highly expressed in airway epithelial cells, while SULF1 mRNA is predominantly expressed in the pulmonary interstitium, outside the airway epithelium (Fig. 7A). In vitro cell experiments further confirmed that SULF1 mRNA is highly expressed in huamn lung fibroblasts (HFL1) (Fig. 7B). HFL1 were stimulated with cigarette smoke extract (CSE), and cell viability was assessed using a CCK-8 assay to determine the appropriate concentration for subsequent experiments. A concentration of 1% CSE, which did not markedly affect cell viability, was selected (Figure S3D). Western blot analysis showed that SULF1 protein expression was significantly decreased after CSE stimulation (Figure S3E), whereas ELISA revealed increased SULF1 protein levels in the culture supernatant (Figure S3F), which consistent with the trend observed in lung tissues and BALF/serum of cigarette smoke–exposed mice. Fluorescence co-localization experiments demonstrated that SULF1 co-localizes with the airway epithelial cell marker E-cadherin (CDH1), but not with the lung fibroblast marker Collagen Type III Alpha 1 Chain (COL3A1) (Fig. 7C). This suggests that airway epithelial cells are the primary target cells through which SULF1 exerts its function in lung tissue.

Fig. 7.

SULF1 and keratin-15 (KRT15) colocalize in airway epithelial cells of CS-exposed mice and physically interact in lung tissue. A RNA-FISH/IF detection of SULF1 mRNA (red) and SULF1 protein (green), with DAPI nuclear staining (blue), imaged at 20X and 60X. B RT-qPCR analysis of SULF1 mRNA expression in THP-1, airway epithelium, 16HBE, BEAS-2B, HFL1, and HBF cells. C IF colocalization of SULF1 (green) with COL3A1 (red) and CDH1 (yellow) in lung sections (20X). D IF staining for KRT15-positive cells in WT and CS mice, imaged at 20X and 60X(n = 3). E Dual-positive SULF1+/KRT15+ cell colocalization in WT and CS mice using IF staining, imaged at 20X and 60X (n = 3). F, G Western blot detection of KRT15 in AIR-WT, AIR-KO,CS-WT, and CS-KO groups(n=6). (H) Co-immunoprecipitation analysis of SULF1-KRT15 interaction using anti-FLAG antibody in lung lysates from SULF1-AAV-overexpressing CS-exposed mice. I Proposed mechanism of SULF1-mediated protection in COPD, Diagram created with Figdraw (Export ID: QIIYOfec48). Statistical analysis was performed using Student’s t-test for panels D and (E), and one-way ANOVA followed by Bonferroni post-hoc testing for panels (G).

Proteomic sequencing was performed on lung tissues from four wild-type (CS-WT) and three SULF1 knockout (CS-KO) mice exposed to CS for 6 months. Principal component analysis (PCA) revealed distinct clustering of protein expression profiles between the two groups (Figure S2A). Among the differentially expressed proteins, a set of keratins, including KRT10, KRT1, KRT6a, KRT78, KRT2, KRT76, KRT15, KRT42, KRT79, KRT5, and KRT17, were consistently upregulated in the CS-KO group (Figures S2B, S2C). Functional enrichment analysis indicated that these upregulated keratins were primarily involved in intermediate filament organization and keratinization (Figure S2D). Previous studies have reported that KRT15 plays a protective role in alveolar integrity [34]. Immunofluorescence staining showed that KRT15 is mainly expressed in airway epithelial cells, and its expression intensity was significantly enhanced after CS exposure (Fig. 7D). Double-staining immunofluorescence experiments for SULF1 and KRT15 indicated co-localization of both proteins in the airway epithelium. The proportion of SULF1+/KRT15 + double-positive cells was significantly increased in the lung tissues of CS-exposed mice (Fig. 7E). qPCR analysis revealed that, compared with the AIR-WT group, SULF1 mRNA levels were significantly decreased in the AIR-KO group, significantly increased in the CS-WT group, and further decreased in the CS-KO group (Figure S3A). Western blot analysis showed that SULF1 protein expression was significantly reduced in the AIR-KO and CS-WT groups relative to the AIR-WT group, and was further decreased in the CS-KO group compared with the CS-WT group (Figure S3B).Western blot analysis showed that, compared to the AIR-WT group, KRT15 levels were upregulated in CS-WT mice (p = 0.06), and downregulated in AIR-KO mice, though these were not statistically significant. However, compared to the AIR-KO group, KRT15 expression was significantly upregulated in CS-KO mice (Fig. 7F, G). Co-immunoprecipitation (Co-IP) assays further suggested an interaction between SULF1 and KRT15 in mouse lung tissue (Fig. 7H).

Discussion

COPD is a complex and heterogeneous respiratory disorder characterized by partially irreversible airflow limitation [35, 36]. Its pathogenesis involves the participation and interaction of multiple pulmonary structural and inflammatory cell types, ultimately leading to persistent airway inflammation and structural remodeling [37, 38]. Traditionally, airway epithelial cells have been the major focus of COPD research [39]; however, increasing evidence has highlighted the critical role of lung fibroblasts in COPD pathophysiology in recent years [40, 41]. Therefore, identifying key fibroblast-derived biomarkers and elucidating their functional roles and mechanisms will provide novel insights for COPD management.

In this study, by integrating single-cell RNA sequencing data with human cohort transcriptomic datasets, we identified SULF1 as a fibroblast-derived core gene associated with lung function decline in COPD. We quantified SULF1 levels in human serum, as well as in mouse serum and BALF, using ELISA. We further generated SULF1-KO mice via CRISPR technology and induced SULF1 overexpression in vivo using AAV, and a CS-induced COPD mouse model was established. Proteomic profiling identified KRT15 as a differentially expressed protein between CS-exposed SULF1-KO and control mice. Our findings suggest that SULF1 may exert protective effects in COPD through systemic compensation and interaction with KRT15 in lung tissue. Specifically, fibroblast-derived SULF1 may enter the serum and BALF as a compensatory response to cigarette smoke-induced lung injury, while its interaction with KRT15 in lung tissue may further contribute to its protective role during COPD progression. In addition, SULF1 may contribute to local fibroblast–epithelial communication, potentially through a paracrine mechanism.

SULF1 is an important extracellular heparan sulfate 6-O-endosulfatase that modifies the sulfation pattern of HSPGs chains, thereby regulating downstream signaling pathways involved in multiple biological processes [42]. Research on SULF1 in pulmonary diseases is still limited. Although SULF1 has been identified as a COPD-related hub gene [31], its mechanistic role has not been previously reported. Notably, genes involved in HSPG sulfation, such as HS6ST1 and SUMF1, have been shown to contribute to alveolar homeostasis [43–45], suggesting that SULF1 may also hold mechanistic significance in COPD development.

We found that serum SULF1 levels were significantly elevated in COPD patients, showing a negative correlation with the FEV₁/FVC ratio and annual FEV₁/FVC ratio decline. Additionally, SULF1 expression was decreased in lung tissue but increased in serum and BALF in CS-exposed mice. We therefore hypothesize that SULF1 may exert protection in COPD partly through a possible compensatory response into the systemic circulation to counteract functional deterioration.

To further validate SULF1 function, we conducted CRISPR-mediated KO and AAV-mediated overexpression experiments in mice. Consistent with our hypothesis, SULF1 deletion significantly exacerbated CS-induced alveolar destruction and lung function impairment, whereas SULF1 overexpression alleviated these pathological alterations, confirming its protective role in COPD. COPD is increasingly recognized as a systemic disease rather than a disease confined to the lungs [46, 47]. Correspondingly, our findings support a systemic regulatory role of SULF1 in COPD progression. Elevated SULF1 levels in both human and mouse peripheral circulation suggest a compensatory protective response to lung injury. Although Yang et al. reported decreased serum SULF1 in COPD patients [48], the discrepancy may be attributed to differences in cohort selection, disease heterogeneity, and disease stage. Importantly, in our clinical cohort, higher baseline serum SULF1 levels were associated with a lower annual decline in FEV1/FVC, suggesting that elevated baseline circulating SULF1 may be related to delayed longitudinal deterioration in lung function. Although serum SULF1 levels showed a negative correlation with baseline FEV1/FVC, this observation may reflect a compensatory increase in SULF1 in the context of established COPD-related lung injury. Additionally, the discordance between lung tissue and peripheral SULF1 protein levels in CS-exposed mice may reflect an altered tissue distribution of SULF1 and suggests a possible extracellular or circulating compensatory response to cigarette smoke-induced lung injury. However, this underlying mechanism remain to be further clarified.

We further observed increased SULF1 mRNA expression in both CS-exposed mice and COPD patient lung tissue, while SULF1 protein was decreased in mouse lung tissue. RNA-FISH and immunofluorescence staining revealed that SULF1 mRNA was predominantly expressed in fibroblasts rather than airway epithelial cells, whereas SULF1 protein accumulated mainly in airway epithelial cells. Combined with its detectable presence in serum and BALF, these results support the secreted nature of SULF1 and its potential paracrine properties. Paracrine regulation is recognized as a crucial cell-cell communication pathway in COPD pathophysiology [49]. Thus, SULF1 protein detected in airway epithelial cells is likely fibroblast-derived, highlighting a potential fibroblast–epithelium signaling axis, although further studies are needed to validate this possibility.

Additionally, our findings identify KRT15 as an interacting protein of SULF1. KRT15, a type I keratin expressed in epithelial cells, is essential for cytoskeletal stability and tissue integrity [32, 50], and has been shown to exert protective roles in COPD [34]. In our study, SULF1 and KRT15 showed clear colocalization in airway epithelial cells. Moreover, under CS exposure, SULF1 protein was downregulated while KRT15 was upregulated, and KRT15 expression was further increased in SULF1-KO mice, suggesting that SULF1 loss may impair epithelial defense capacity and trigger compensatory KRT15 upregulation. We therefore propose that fibroblast-derived SULF1 is secreted via paracrine transport to airway epithelial cells, where it interacts with and promotes KRT15 expression to confer cooperative protection. However, although our data support the colocalization of SULF1 and KRT15 in the airway epithelial region and their interaction at the lung tissue level, they do not directly demonstrate that SULF1 interacts with KRT15 within airway epithelial cells. Future studies using airway epithelial cell-specific co-immunoprecipitation assays are therefore needed to further determine whether SULF1 directly interacts with KRT15 in airway epithelial cells and to clarify the functional relevance of this potential interaction.

In CSE-stimulated HFL1 cells, intracellular SULF1 protein was decreased, whereas supernatant SULF1 was increased, further supporting the possibility that SULF1 protein detected in airway epithelial regions may be at least partly derived from lung fibroblasts. However, the paracrine mechanism involved in fibroblast–epithelial communication remains to be further elucidated. However, without dedicated secretion assays and fibroblast-specific conditional SULF1 knockout mice, the altered secretion dynamics, precise cellular source of circulating SULF1, and proposed paracrine mechanism could not be fully confirmed. Therefore, we were unable to more precisely define the cellular source of circulating SULF1 or fully validate the proposed paracrine mechanism in vivo. In addition, in the present study, although KRT15 was prioritized for further analysis, the potential involvement of other differentially expressed keratins cannot be excluded. Therefore, the SULF1–KRT15 axis may represent an important, but not exclusive, mechanism in COPD. Notably, the current immunoprecipitation data were obtained using a FLAG-tagged SULF1 overexpression system and support an interaction between overexpressed SULF1 and KRT15. However, this interaction remains to be further validated under endogenous or knockout conditions. Although a physical association between SULF1 and KRT15 was observed, the molecular consequence of this interaction remains to be further defined. Thus, the SULF1–KRT15 axis should be interpreted as a potential functional association rather than a fully established causal mechanism. Future rescue experiments in airway epithelial cells are needed to clarify its directionality and downstream effects.

Moreover, the molecular mechanism underlying the interaction between SULF1 and KRT15 in epithelial cells warrants further investigation, which will help elucidate the precise protective mechanism of SULF1 in COPD.

Conclusion

In conclusion, our study is the first to demonstrate that SULF1 exerts a protective role in COPD progression. CS-induced SULF1-KO mice exhibited worsened pulmonary pathology and lung function, whereas AAV-mediated SULF1 overexpression significantly alleviated these changes. The underlying mechanism may involve paracrine fibroblast-to-epithelium signaling and SULF1-KRT15 interaction, as well as circulating SULF1-mediated functional compensation (Fig. 7I). However, the clinical utility of SULF1 or the SULF1-KRT15 axis as diagnostic or prognostic biomarkers requires further evaluation. Collectively, these findings suggest that SULF1 holds promise as a novel therapeutic target for precision treatment in COPD.

Authors’ contributions

Author contributions: Study design: YL, EY, PR, and YZ; Experimental procedures: YL, EY, YX, YZ, XW, HW, QW, XM, HL, JZ, and XZ; Clinical data collection: FW, ZW, QW, and GT; Data analyse: YL and EY; Results visualization: YL and EY; Results interpretation: All authors; Manuscript writing: YL; Manuscript revising: YL, EY, PR, and YZ; Project guidance and funding acquisition: XM, RS, PR, and YZ;. All authors approved the final version of the manuscript.

Funding

This study was supported by the Noncommunicable Chronic Diseases-National Science and Technology Major Project (2024ZD0528401), the Foundation of Guangzhou National Laboratory (SRPG22-016 and SRPG22-018), the National Natural Science Foundation of China (82570065), the Major Project of Guangzhou National Laboratory(GZNL2024A02005), National Natural Science Foundation of China (82500053), and the grant of State Key Laboratory of Respiratory Disease (SKLRD-Z-202315).

Data availability

The datasets supporting the conclusions of this article are available in online repositories. Names of repositories are provided in the article/Additional Material. All other relevant materials are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

Peripheral blood samples and clinical information from COPD patients and controls were obtained from the Early Chronic Obstructive Pulmonary Disease (ECOPD) Study (Clinical Trial Registration No. ChiCTR190002464, July 19, 2019). Written informed consent was obtained from all participants.

Consent for publication

All authors have read and approved the final version of the manuscript, and consent to its publication in Respiratory Research.

Competing interests

The authors declare no competing interests.