基于WGCNA和机器学习的特发性肺纤维化和系统性硬化症的常见生物标志物-动脉网

Common biomarkers of idiopathic pulmonary fibrosis and systemic sclerosis based on WGCNA and machine learning

Nature 等信源发布 2025-01-03 16:49



可切换为仅中文







Abstract

摘要

Interstitial lung disease (ILD) is known to be a major complication of systemic sclerosis (SSc) and a leading cause of death in SSc patients. As the most common type of ILD, the pathogenesis of idiopathic pulmonary fibrosis (IPF) has not been fully elucidated. In this study, weighted correlation network analysis (WGCNA), protein‒protein interaction, Kaplan–Meier curve, univariate Cox analysis and machine learning methods were used on datasets from the Gene Expression Omnibus database.

间质性肺病（ILD）是系统性硬化症（SSc）的主要并发症，也是SSc患者死亡的主要原因。作为最常见的ILD类型，特发性肺纤维化（IPF）的发病机制尚未完全阐明。在这项研究中，对Gene Expression Omnibus数据库的数据集使用了加权相关网络分析（WGCNA），蛋白质-蛋白质相互作用，Kaplan-Meier曲线，单变量Cox分析和机器学习方法。

CCL2 was identified as a common characteristic gene of IPF and SSc. The genes associated with CCL2 expression in both diseases were enriched mainly in chemokine-related pathways and lipid metabolism-related pathways according to Gene Set Enrichment Analysis. Single-cell RNA sequencing (sc-RNAseq) revealed a significant difference in CCL2 expression in alveolar epithelial type 1/2 cells, mast cells, ciliated cells, club cells, fibroblasts, M1/M2 macrophages, monocytes and plasma cells between IPF patients and healthy donors.

CCL2被鉴定为IPF和SSc的共同特征基因。根据基因组富集分析，两种疾病中与CCL2表达相关的基因主要富集在趋化因子相关途径和脂质代谢相关途径中。单细胞RNA测序（sc RNAseq）显示IPF患者和健康供体之间肺泡上皮1/2型细胞，肥大细胞，纤毛细胞，俱乐部细胞，成纤维细胞，M1/M2巨噬细胞，单核细胞和浆细胞中CCL2表达的显着差异。

Statistical analyses revealed that CCL2 was negatively correlated with lung function in IPF patients and decreased after mycophenolate mofetil (MMF) treatment in SSc patients. Finally, we identified CCL2 as a common biomarker from IPF and SSc, revealing the common mechanism of these two diseases and providing clues for the study of the treatment and mechanism of these two diseases..

统计分析显示，IPF患者的CCL2与肺功能呈负相关，SSc患者的霉酚酸酯（MMF）治疗后CCL2降低。最后，我们将CCL2鉴定为IPF和SSc的常见生物标志物，揭示了这两种疾病的共同机制，并为研究这两种疾病的治疗和机制提供了线索。。

Introduction

导言

IPF is defined as chronic, progressive ILD that primarily occurs in older people; the median survival time after diagnosis is approximately 2–3 years, and its pathogenesis is still unclear

IPF被定义为主要发生在老年人中的慢性进行性ILD；诊断后的中位生存时间约为2-3年，其发病机制尚不清楚

. The BRIC countries (Brazil, Russia, India, and China) may include 1 million cases of IPF

。金砖四国（巴西、俄罗斯、印度和中国）可能包括100万例IPF

. For IPF management, pirfenidone and the triple tyrosine kinase [vascular endothelial growth factor (VEGF), fibroblast growth factor (FGF) and platelet-derived growth factor (PDGF)] inhibitor nintedanib are recommended to delay the progression of fibrosis, which cannot be reversed

对于IPF管理，建议使用吡非尼酮和三重酪氨酸激酶[血管内皮生长因子（VEGF），成纤维细胞生长因子（FGF）和血小板衍生生长因子（PDGF）]抑制剂nintedanib来延缓纤维化的进展，这是无法逆转的

. Moreover, lung transplantation is considered the final method for patients with moderate to severe disease

此外，肺移植被认为是中重度疾病患者的最终方法

Current studies suggest that the pathogenesis of IPF may involve transforming growth factor-β (TGF-β), which is increased in patients with pulmonary fibrosis

目前的研究表明，IPF的发病机制可能涉及转化生长因子-β（TGF-β），而转化生长因子-β在肺纤维化患者中增加

. VEGF can improve pulmonary hypertension (PH), but simultaneously aggravates pulmonary fibrosis, which may play opposite roles in different lung compartments

VEGF可以改善肺动脉高压（PH），但同时加重肺纤维化，这可能在不同的肺区室中发挥相反的作用

SSc, also known as scleroderma, is a type of autoimmune connective-tissue disease (CTD), with the clinical presentation of hard, thickened areas of skin, Raynaud’s phenomenon, and gastroesophageal reflux, and has the highest mortality rates of all rheumatic diseases

SSc，也称为硬皮病，是一种自身免疫性结缔组织病（CTD），临床表现为皮肤坚硬增厚，雷诺现象和胃食管反流，在所有风湿性疾病中死亡率最高

. The pathogenesis of SSc varies, is complex, has not been fully elucidated, and may be related to genetics, the environment, disorders of the immune response and vascular lesions

SSc的发病机制多种多样，很复杂，尚未完全阐明，可能与遗传学，环境，免疫反应障碍和血管病变有关

. The prevalence in Europe ranges from 7.2 to 33.9 per 100,000 individuals

。欧洲的患病率为每100000人7.2至33.9

. Patients with SSc are often difficult to diagnose early in the course of the disease, delaying their appropriate treatment and care

。SSc患者通常很难在疾病的早期诊断，从而延误了他们的适当治疗和护理

With respect to pathogenesis, some studies have proposed that the expression of TGF-β, TGF-βR1 and TGF-βR2 is increased in the skin cells of SSc patients, indicating that TGF-β may play a role in SSc skin lesions

关于发病机制，一些研究表明，SSc患者皮肤细胞中TGF-β，TGF-βR1和TGF-βR2的表达增加，表明TGF-β可能在SSc皮肤病变中起作用

. The level of VEGF in the plasma of SSc patients increases with increasing clinical stage. It is believed to be involved in the disease progression of SSc and in the remodelling of the skin microvasculature by hypoxia-induced endothelial‒mesenchymal transition (EndoMT)

SSc患者血浆中VEGF水平随着临床分期的增加而增加。据信它与SSc的疾病进展以及缺氧诱导的内皮-间质转化（EndoMT）对皮肤微血管的重塑有关

. For the management of SSc patients, methotrexate and MMF are recommended for skin-associated complications

对于SSc患者的治疗，建议使用甲氨蝶呤和MMF治疗皮肤相关并发症

The pathogenesis and metabolic pathways of IPF and SSc are related. CCL2, a chemokine, has been shown to be involved in the pathogenesis of IPF by being regulated by the FOXF1/R-Ras signalling, NFATc3 and in the pathogenesis of SSc by being a downstream signalling molecule with low NCF1 activity

。趋化因子CCL2已被证明通过受FOXF1/R-Ras信号传导，NFATc3的调节而参与IPF的发病机制，并且通过作为具有低NCF1活性的下游信号分子而参与SSc的发病机制

. The pathogenesis of these two diseases has not been fully elucidated, and existing treatments are not curable, placing a large burden on patients and health care systems. Therefore, the need to study the common markers and molecular pathways of IPF and SSc is urgent. In this study, WGCNA and machine learning were used to select markers related to the disease status of these 2 diseases and the prognosis of IPF, and the markers were then analysed via multiple methods, providing a possible direction for the pathogenesis and treatment of IPF and SSc..

。这两种疾病的发病机制尚未完全阐明，现有的治疗方法无法治愈，给患者和医疗保健系统带来了沉重负担。因此，迫切需要研究IPF和SSc的常见标记和分子途径。在这项研究中，使用WGCNA和机器学习来选择与这两种疾病的疾病状态和IPF预后相关的标记，然后通过多种方法分析这些标记，为IPF和SSc的发病机制和治疗提供了可能的方向。。

Results

结果

Removal of batch effects and PCA

去除批次效应和PCA

The research flowchart of this research was shown in Fig.

这项研究的研究流程图如图所示。

. Batch effects were eliminated between the 3 cohorts in GSE70866 (Fig.

GSE70866中的3个队列之间消除了批次效应（图）。

a,b) and the baseline IPF samples in GSE27957, GSE28042, and GSE93606 (Fig.

a、 b）和GSE27957、GSE28042和GSE93606中的基线IPF样本（图）。

c,d).

c、 d）。

Fig. 1

图1

Flow chart of the study.

研究流程图。

Full size image

全尺寸图像

Fig. 2

图2

Principal Component Analysis (PCA) of IPF datasets before and after batch correction and normalization. (

批量校正和归一化前后IPF数据集的主成分分析（PCA）。(

一

b类

) PCA of 3 cohorts in GSE70866 before and after batch correction and normalization. (

)在批次校正和归一化之前和之后，GSE70866中3个队列的PCA。(

c级

) PCA of baseline IPF samples in GSE27957, GSE28042 and GSE93606 before and after batch correction and normalization.

)在批次校正和归一化之前和之后，GSE27957，GSE28042和GSE93606中基线IPF样品的PCA。

Full size image

全尺寸图像

WGCNA, PPI, survival analysis and machine learning were used to identify CCL2 as a common characteristic gene for IPF and SSc diagnosis and the prognosis of IPF

使用WGCNA，PPI，生存分析和机器学习将CCL2鉴定为IPF和SSc诊断以及IPF预后的共同特征基因

As shown in Fig.

如图所示。

a–h, we performed WGCNA for GSE70866 and GSE181549 to identify the module most strongly correlated with IPF and SSc. Figure

a–h，我们对GSE70866和GSE181549进行了WGCNA，以确定与IPF和SSc最密切相关的模块。图

a,c,e,f are for GSE70866, Fig.

a、 c，e，f代表GSE70866，图。

b,d,g,h. are for GSE181549.

b、 d、g、h代表GSE181549。

Fig. 3

图3

Weighted gene co-expression network analysis for identification and analysis of characteristic genes common to the IPF and SSc datasets. (

加权基因共表达网络分析，用于鉴定和分析IPF和SSc数据集共有的特征基因。(

一

b类

) The left panel shows the scale-free fitting exponent analysis with multiple soft threshold powers (β), and the right panel shows the average connectivity analysis with multiple soft threshold powers. Figure a is for the IPF dataset, and figure b is for the SSc dataset. (

)左图显示了具有多个软阈值功率（β）的无标度拟合指数分析，右图显示了具有多个软阈值功率的平均连通性分析。图a是IPF数据集，图b是SSc数据集。(

c级

) Heatmap showing the relationships between module eigengenes and IPF or SSc status. The correlation (left) and p-value (right) of the module feature genes with disease status are shown. Figure c shows the IPF dataset, and figure d shows the SSc dataset. (

)热图显示模块特征基因与IPF或SSc状态之间的关系。显示了模块特征基因与疾病状态的相关性（左）和p值（右）。图c显示了IPF数据集，图d显示了SSc数据集。(

–

小时

) The correlation plot between the module membership and the gene significance of genes in each module. (

)模块成员与每个模块中基因的基因重要性之间的相关图。(

f级

) correspond to the darkgreen and magenta modules in the IPF dataset, respectively, (

)分别对应于IPF数据集中的暗绿色和品红色模块(

克

小时

) correspond to the salmon and turquoise modules in the SSc dataset, respectively. (

)分别对应于SSc数据集中的鲑鱼和绿松石模块。(

我

) The intersection of genes in key modules of two diseases was obtained via a Venn diagram. (

)通过维恩图获得了两种疾病关键模块中基因的交集。(

) Based on the MCC method of the Cytoscape plug-in CytoHubba, PPI network analysis was performed for the top 10 genes among the 138 intersecting genes.

)。

Full size image

全尺寸图像

For GSE70866, β = 6 (scale-free R

对于GSE70866，β=6（无标度R）

= 0.85) was chosen as the “soft” threshold (Fig.

=（0.85）被选为“软”阈值（图）。

a). Then, according to Fig.

a）。然后，根据图。

c, we chose darkgreen module and magenta module as the modules that were most strongly correlated with the IPF. In addition, we found that there was a strong association between module membership and gene significance in these 2 modules (Fig.

c、我们选择暗绿色模块和洋红色模块作为与IPF最密切相关的模块。此外，我们发现在这两个模块中，模块成员与基因重要性之间存在很强的关联（图）。

e,f).

e、 f）。

For GSE181549, we chose β = 12 (scale-free R

对于GSE181549，我们选择了12（无标度R

= 0.85) as the “soft” threshold (Fig.

=0.85）作为“软”阈值（图）。

b). On the basis of Fig.

b）。在图的基础上。

d, the salmon module and turquoise module were chosen as the modules that were most strongly correlated with SSc. Moreover, a strong association was found between module membership and gene significance in these 2 modules (Fig.

d、选择鲑鱼模块和绿松石模块作为与SSc相关性最强的模块。此外，在这两个模块中，模块成员与基因重要性之间存在很强的关联（图）。

g,h).

g、 h）。

The intersection of the above two pairs of most correlated modules contained 138 genes (Fig.

上述两对最相关模块的交集包含138个基因（图）。

i). To determine the relationship between potential pathogenic genes common to IPF and SSc, the 138 intersection genes were input into the STRING database (

i）。为了确定IPF和SSc共有的潜在致病基因之间的关系，将138个交叉基因输入STRING数据库(

https://www.string-db.org/

) with a medium confidence score of > 0.4. These common potential pathogenic genes were visualized via Cytoscape software, and the top rank 10 genes were identified via maximum clique centrality (MCC) methods (Fig.

)中等置信度得分为>0.4。通过Cytoscape软件可视化这些常见的潜在致病基因，并通过最大集团中心性（MCC）方法鉴定排名前10位的基因（图）。

j).

j）。

K–M curves of these 10 genes were plotted on the basis of the prognostic information of the IPF samples from GSE70866, which revealed that all 10 genes were risk factors for the prognosis of IPF (Fig.

根据来自GSE70866的IPF样本的预后信息绘制了这10个基因的K-M曲线，这表明所有10个基因都是IPF预后的危险因素（图）。

a–j). Univariate Cox regression revealed that the hazard ratios (HRs) of these 10 genes were all greater than 1, indicating that these 10 genes are risk factors for the prognosis of IPF (Fig.

a–j）。单因素Cox回归分析显示，这10个基因的危险比（HR）均大于1，表明这10个基因是IPF预后的危险因素（图）。

k).

k）。

Fig. 4

图4

Survival analysis of the influence of 10 intersecting genes on the prognosis of IPF. (

10个交叉基因对IPF预后影响的生存分析。(

一

–

) K–M curves of the influence of 10 intersecting genes on the prognosis of IPF, a to j represent CCL2, CXCL1, ICAM1, CCL3, CXCR4, MMP9, TLR2, CCL4, MMP1 and TIMP1, respectively. (

)10个相交基因对IPF预后影响的K–M曲线，a至j分别代表CCL2，CXCL1，ICAM1，CCL3，CXCR4，MMP9，TLR2，CCL4，MMP1和TIMP1。(

) Forest plot of univariate Cox analysis of the effects of 10 intersecting genes on IPF prognosis.

)。

Full size image

全尺寸图像

Several machine learning algorithms were then applied to both the merged dataset of baseline IPF samples in GSE70866 and the SSc samples in GSE58095. As shown in Fig.

然后将几种机器学习算法应用于GSE70866中基线IPF样本的合并数据集和GSE58095中的SSc样本。如图所示。

a,b, Least Absolute Shrinkage and Selection Operator (LASSO)-cox algorithm identified 5 potential biomarkers, and as illustrated in Fig.

a、 b，最小绝对收缩和选择算子（LASSO）-cox算法确定了5种潜在的生物标志物，如图所示。

c, Support Vector Machines-Recursive Feature Elimination (SVM-RFE) algorithm revealed that a model involving 1 gene achieved the lowest Root Mean Square Error (RMSE). Moreover, 3 genes were identified via the random forest algorithm (Fig.

c、支持向量机递归特征消除（SVM-RFE）算法显示，涉及1个基因的模型实现了最低的均方根误差（RMSE）。此外，通过随机森林算法鉴定了3个基因（图）。

d). By intersecting the results of these 3 algorithms, we defined CCL2 as the common characteristic gene in IPF prognosis. As shown in Fig.

d）。通过相交这三种算法的结果，我们将CCL2定义为IPF预后的共同特征基因。如图所示。

e,f, the LASSO algorithm defines 3 biomarkers associated with SSc diagnosis. Figure

e、 f，LASSO算法定义了3种与SSc诊断相关的生物标志物。图

g shows that SVM-RFE indicates that the 4-gene model has the lowest RMSE. In addition, the random forest results revealed that 2 biomarkers had higher Mean Decrease Gini values (Fig.

g表明SVM-RFE表明4基因模型的RMSE最低。此外，随机森林结果显示，2个生物标志物具有较高的平均降低基尼值（图）。

h). From the intersection of the results of these 3 machine learning algorithms, we selected TIMP1 and CCL2 as biomarkers for SSc diagnosis. In summary, CCL2 was selected as the biomarker most strongly associated with IPF prognosis and the diagnosis of both IPF and SSc.

h）。从这3种机器学习算法的结果的交叉点，我们选择TIMP1和CCL2作为SSc诊断的生物标志物。总之，CCL2被选为与IPF预后以及IPF和SSc诊断最密切相关的生物标志物。

Fig. 5

图5

Machine learning was used to screen characteristic genes for IPF prognosis and SSc diagnosis. (

机器学习用于筛选IPF预后和SSc诊断的特征基因。(

一

b类

) Biomarker screening in LASSO-cox models. The number of genes corresponding to the lowest point of the curve (n = 5) was most suitable for assessing the prognosis of IPF. (

)LASSO-cox模型中的生物标志物筛选。对应于曲线最低点的基因数量（n=5）最适合评估IPF的预后。(

c级

克

) Screening for biomarkers via the SVM-RFE algorithm. (

)通过SVM-RFE算法筛选生物标志物。(

小时

): The random forest method was used to screen biomarkers, sequencing the importance of genes according to MeanDecreaseGini values. (

)：使用随机森林方法筛选生物标志物，根据MeanDecreaseGini值对基因的重要性进行测序。(

f级

) Biomarker screening in LASSO models. The number of genes corresponding to the lowest point of the curve (n = 3) was most suitable for assessing the diagnosis of SSc.

)LASSO模型中的生物标志物筛选。对应于曲线最低点的基因数量（n=3）最适合评估SSc的诊断。

Full size image

全尺寸图像

GSEA enrichment analysis

GSEA富集分析

The genes up-regulated with IPF patients’ CCL2 expression value was mainly enriched in immune related components, especially the chemokine family (Fig.

IPF患者CCL2表达值上调的基因主要富含免疫相关成分，尤其是趋化因子家族（图）。

a,c), such as positive regulation of chemokine production, CCR chemokine receptor binding, chemokine activity, chemokine receptor binding, CXCR chemokine receptor binding and cytokine receptor binding. Other components and pathways that are positively associated with CCL2 granule in IPF patients include azurophil granule and several immune-related pathways, such as allograft rejection, asthma, graft-versus-host disease and viral protein interaction with cytokine and cytokine receptor (Fig. .

a、 c），如趋化因子产生的正调控，CCR趋化因子受体结合，趋化因子活性，趋化因子受体结合，CXCR趋化因子受体结合和细胞因子受体结合。IPF患者中与CCL2颗粒正相关的其他成分和途径包括嗜天青颗粒和几种免疫相关途径，例如同种异体移植排斥反应，哮喘，移植物抗宿主病以及病毒蛋白与细胞因子和细胞因子受体的相互作用（图。

b,d).

b、 d）。

Fig. 6

图6

GSEA based on the CCL2 expression level. (

GSEA基于CCL2表达水平。(

一

–

) are for the IPF dataset, (

)用于IPF数据集(

–

小时

) are for the SSc dataset. (

)用于SSc数据集。(

一

) GO: BP analysis via GSEA. (

)GO：通过GSEA进行BP分析。(

b类

f级

) GO: CC analysis via GSEA. (

)GO：通过GSEA进行CC分析。(

c级

克

) GO: MF analysis via GSEA. (

)GO：通过GSEA进行MF分析。(

小时

) KEGG analysis via GSEA.

)通过GSEA进行KEGG分析。

Full size image

全尺寸图像

In the SSc dataset, the gene up-regulated with CCL2 expression value was enriched mainly in lipid metabolism, such as acyl − CoA metabolic process, cholesterol biosynthetic process, establishment of skin barrier, fatty − acyl − CoA biosynthetic process, fatty acid derivative biosynthetic process, fatty acid derivative metabolic process, mitochondrial protein-containing complex, fatty acid synthase activity, biosynthesis of unsaturated fatty acids, butanoate metabolism, fatty acid degradation, fatty acid elongation and fatty acid metabolism (Fig. .

在SSc数据集中，CCL2表达值上调的基因主要富集在脂质代谢中，如酰基辅酶A代谢过程、胆固醇生物合成过程、皮肤屏障的建立、脂肪酰基辅酶A生物合成过程、脂肪酸衍生物生物合成过程、脂肪酸衍生物代谢过程、线粒体蛋白复合物、脂肪酸合成酶活性、不饱和脂肪酸生物合成、丁酸代谢、脂肪酸降解、脂肪酸伸长和脂肪酸代谢（图）。

e–h). We found that these genes are involved mainly in fatty acid metabolism.

e–h）。我们发现这些基因主要参与脂肪酸代谢。

Common characteristic gene’s expression in single cells

共同特征基因在单细胞中的表达

4 IPF samples and 4 healthy samples were chosen for single-cell analysis (Fig.

选择4个IPF样品和4个健康样品进行单细胞分析（图）。

a). All the cells in these 8 samples’ UMAP were annotated into 14 categories: Alveolar Epithelial Type 1 Cell [number of cells in IPF group (n1) = 250, number of cells in Healthy Donor group (n2) = 553], Alveolar Epithelial Type 2 Cell (n1 = 2168,n2 = 10,089), B Cell (n1 = 412,n2 = 11), Mast Cell (n1 = 137,n2 = 81), Ciliated Cell (n1 = 397, n2 = 209), Club Cell (n1 = 596,n2 = 226), Fibroblast (n1 = 137,n2 = 135), Lymphatic Cell (n1 = 35,n2 = 86), M1 Macrophage (n1 = 3183,n2 = 1645), M2 Macrophage (n1 = 2212,n2 = 5230), Monocyte (n1 = 1424,n2 = 537), Plasma Cell (n1 = 886,n2 = 146), T Cell (n1 = 611,n2 = 35), Vascular Endothelial Cell (n1 = 245,n2 = 257) (Fig. .

一这8个样本的UMAP中的所有细胞都被注释为14类：肺泡上皮1型细胞Medendeendendendenden和eenden和teenden和ween和eeneden和eendeen和eenedeen和eenedee和eenedeenden。二、二、三、四、五内皮细胞（n1=886，n2塌陷146）、T细胞（n1塌陷611，n2塌陷35）、血管内皮细胞（n1=245，n2塌陷257）（图塌陷）。

b,c). The expression values of the common characteristic gene CCL2 in IPF and healthy samples were compared (Fig.

b、 c）。比较IPF和健康样品中共同特征基因CCL2的表达值（图）。

d,e), and Mann–Whitney U test was conducted based on the fact that the expression data of all cell types did not follow the normal distribution. Then we found significant differences in CCL2 expression in Alveolar Epithelial Type 1 Cell (

d、 e），Mann-Whitney U检验是基于所有细胞类型的表达数据均不遵循正态分布的事实进行的。然后我们发现肺泡上皮1型细胞中CCL2表达存在显着差异(

< 0.0001) , Alveolar Epithelial Type 2 Cell (

<0.0001），肺泡上皮2型细胞(

< 0.0001), Mast Cell (

<0.0001），肥大细胞(

= 0.0088), Ciliated Cell (

=0.0088），纤毛细胞(

= 0.0023), Club Cell (

=0.0023），俱乐部牢房(

< 0.0001), Fibroblast (

<0.0001），成纤维细胞(

= 0.0001), M1 Macrophage (

=0.0001），M1巨噬细胞(

< 0.0001), M2 Macrophage (

<0.0001），M2巨噬细胞(

< 0.0001), Monocyte (

<0.0001），单核细胞(

< 0.0001) and Plasma Cell (

0.0001）和浆细胞(

< 0.0001), and the tests were all two-tailed (Fig.

<0.0001），测试都是双尾的（图）。

f–s). Because the n2 value of B cells was too small (n2 = 11) and CCL2 expression was 0 in the healthy donor group, the Mann–Whitney U test did not yield a p value and was not included in subsequent analyses.

。由于健康供体组中B细胞的n2值太小（n2=11），CCL2表达为0，因此Mann-Whitney U检验未产生p值，因此不包括在随后的分析中。

Fig. 7

图7

The expression profile of common characteristic genes in IPF single cells and the difference in CCL2 gene expression between IPF patients and healthy controls. (

IPF单细胞中常见特征基因的表达谱以及IPF患者与健康对照之间CCL2基因表达的差异。(

一

) UMAP of 8 samples after removal of the batch effect. (

)去除批次效应后的8个样品的UMAP。(

b类

) Cellular subtypes of IPF lungs. (

)。(

c级

) Cellular subtypes of healthy donor lungs. (

)健康供体肺的细胞亚型。(

) CCL2 gene expression in IPF lungs. (

)IPF肺中CCL2基因的表达。(

) CCL2 gene expression in healthy donor lungs. (

)CCL2基因在健康供体肺中的表达。(

f级

–

) Histogram of the mean expression value of CCL2 in each cell subtype of IPF and healthy control samples, with the standard deviation (SD). *

)IPF和健康对照样品的每个细胞亚型中CCL2的平均表达值的直方图，具有标准偏差（SD）*

< 0.05; **

< 0.01; ***

< 0.001; ****

< 0.0001; ns not significant.

<0.0001；ns不重要。

Full size image

全尺寸图像

CCL2 expression was negatively correlated with lung function in patients with IPF and decreased after MMF treatment in diffuse cutaneous systemic sclerosis (dcSSc) patients

IPF患者CCL2表达与肺功能呈负相关，弥漫性皮肤系统性硬化症（dcSSc）患者MMF治疗后CCL2表达降低

The relationship between IPF patients’ CCL2 expression and their lung function [diffusion capacity of carbon monoxide in the lung (DLco) % predicted and forced vital capacity (FVC) % predicted)] according to GSE32537 is shown in Fig.

根据GSE32537，IPF患者CCL2表达与肺功能[一氧化碳在肺中的扩散能力（DLco）%预测和用力肺活量（FVC）%预测]之间的关系如图所示。

a,b. With increasing CCL2 expression value, DLco % predicted and FVC % predicted decreased in IPF patients. A Wilcoxon matched-pairs signed rank test was performed on the dcSSc samples before and after the use of MMF in GSE97248 (n = 18 in each group). After MMF treatment for 3 months, CCL2 expression was significantly lower than before (.

a、 b.随着CCL2表达值的增加，IPF患者的DLco%预测值和FVC%预测值降低。在GSE97248中使用MMF之前和之后，对dcSSc样品进行Wilcoxon配对符号秩检验（每组n=18）。MMF治疗3个月后，CCL2表达明显低于治疗前（。

= 0.0133) (Fig.

=0.0133）（图。

c).

c）。

Fig. 8

图8

Scatterplot of the relationship between lung CCL2 expression and lung function in patients with IPF. (

IPF患者肺CCL2表达与肺功能关系的散点图。(

一

) Relationship between CCL2 expression and DLco % predicted in lung tissue. (

)CCL2表达与肺组织中预测的DLco%之间的关系。(

b类

) Relationship between CCL2 expression and FVC % predicted in lung tissue. (

)CCL2表达与肺组织中预测的FVC%之间的关系。(

c级

) Box plot of CCL2 expression in the skin tissue of SSc patients before and after treatment with MMF. *

)MMF治疗前后SSc患者皮肤组织中CCL2表达的箱形图*

< 0.05.

Full size image

全尺寸图像

Discussion

讨论

As a systematic biological method, WGCNA can reveal the correlation patterns of gene expression between different samples and identify gene co-expression modules related to phenotypes. Compared with differentially expressed genes, WGCNA not only studies the relationships between genes and phenotypes but also focuses on the interactions between genes, providing an important method for screening disease-related genes.

作为一种系统的生物学方法，WGCNA可以揭示不同样品之间基因表达的相关模式，并鉴定与表型相关的基因共表达模块。与差异表达基因相比，WGCNA不仅研究基因与表型之间的关系，而且关注基因之间的相互作用，为筛选疾病相关基因提供了重要方法。

In our study, the WGCNA algorithm was used to identify the co-expressed modules in IPF and SSc, and then, in order to identify potential closely interacting protein-coding genes in the intersection, thereby screening genes that play more important roles in IPF and SSc, PPI was constructed on the basis of the 138 intersection genes.

在我们的研究中，使用WGCNA算法来鉴定IPF和SSc中的共表达模块，然后，为了鉴定交叉点中潜在的紧密相互作用的蛋白质编码基因，从而筛选在IPF和SSc中起更重要作用的基因，在138个交叉基因的基础上构建了PPI。

To more visually examine and describe the relationship between the incidence of endpoint events (deaths) and the corresponding gene expression levels in IPF patients, we performed survival analyses, including K–M curves and univariate Cox analysis. After survival analysis and machine learning for the 10 genes in the PPI network, the CCL2 gene was identified as an important biomarker that plays a key role in the pathogenesis of IPF and SSc..

为了更直观地检查和描述IPF患者终点事件（死亡）发生率与相应基因表达水平之间的关系，我们进行了生存分析，包括K–M曲线和单变量Cox分析。在对PPI网络中的10个基因进行生存分析和机器学习后，CCL2基因被确定为重要的生物标志物，在IPF和SSc的发病机理中起着关键作用。。

CCL2, also known as monocyte chemoattractant protein-1 (MCP-1), was the first CC chemokine to be discovered and studied; it preferentially binds to its receptor, CCR2, a G-protein-coupled 7 transmembrane receptor

CCL2，也称为单核细胞趋化蛋白-1（MCP-1），是第一个被发现和研究的CC趋化因子；它优先与其受体CCR2结合，CCR2是一种G蛋白偶联的7跨膜受体

. The TGF-β pathway plays an important role in almost all types of fibrosis and involves multiple signalling cascades

TGF-β途径在几乎所有类型的纤维化中都起着重要作用，并涉及多个信号级联

. A previous study revealed that TGF-β1 induces fibroblasts to produce CCL2, which can then induce further fibrotic responses in these cells, our result in Fig.

先前的一项研究表明，TGF-β1诱导成纤维细胞产生CCL2，然后可以在这些细胞中诱导进一步的纤维化反应，我们的结果见图。

l also shows that compared with healthy control, CCL2 expression level is higher in the fibroblast of IPF patients, indicating that the TGF-β pathway might be activated in IPF patients

l还表明，与健康对照组相比，IPF患者成纤维细胞中CCL2的表达水平更高，表明IPF患者中TGF-β途径可能被激活

. The study also revealed that CCL2 is downstream of TGF-β1-induced gene expression, which indicates that CCL2 is a member of the TGF-β pathway

该研究还表明，CCL2位于TGF-β1诱导的基因表达的下游，这表明CCL2是TGF-β途径的成员

. Paôline reported that in SSc patients, CCL2 can be produced at high levels by DC-SIGN- positive alternatively activated macrophages, a type of macrophage related to the degree of skin fibrosis in SSc patients, indicating that CCL2 is involved in the vicious cycle of skin fibrosis in SSc patients

.Paôline报道，在SSc患者中，DC-SIGN阳性交替激活的巨噬细胞可以高水平产生CCL2，这是一种与SSc患者皮肤纤维化程度相关的巨噬细胞，表明CCL2参与了SSc患者皮肤纤维化的恶性循环

. Another study revealed that the serum level of CCL2 in SSc patients is significantly greater than that in controls and is positively correlated with the modified Rodnan skin score (mRSS), confirming the role of CCL2 in extracellular matrix (ECM) deposition

另一项研究表明，SSc患者的血清CCL2水平显着高于对照组，并且与改良的Rodnan皮肤评分（mRSS）呈正相关，证实了CCL2在细胞外基质（ECM）沉积中的作用

Karman showed that in the myeloid-enriched IPF subset, the monocyte‒macrophage chemoattractant axis, which potentially includes the CCL2‒CCR2 axis, was highly activated, and this axis might play a role in recruiting inflammatory macrophages to this subset

卡曼表明，在富含骨髓的IPF亚群中，单核细胞-巨噬细胞趋化轴（可能包括CCL2-CCR2轴）被高度激活，该轴可能在向该亚群募集炎性巨噬细胞中发挥作用

. Our GSEA results revealed that the genes upregulated in IPF patients with high CCL2 expression were enriched in monocyte chemotaxis, positive regulation of chemokine production, CCR chemokine receptor binding, chemokine activity and chemokine receptor binding. We speculate that in patients with IPF, the chemokine system is widely activated and might be a target for the treatment of IPF..

我们的GSEA结果显示，在具有高CCL2表达的IPF患者中上调的基因富含单核细胞趋化性，趋化因子产生的正调节，CCR趋化因子受体结合，趋化因子活性和趋化因子受体结合。我们推测，在IPF患者中，趋化因子系统被广泛激活，可能是治疗IPF的靶点。。

Chemokines were initially found to be mediators of the targeted migration of immune cells to sites of inflammation and injury, including 4 major subfamilies, CC chemokines, CXC chemokines, CX3C chemokines and XC chemokines, and each of these subfamilies has its own receptor: CCRs, CXCRs, CX3CRs and XCRs.

趋化因子最初被发现是免疫细胞向炎症和损伤部位靶向迁移的介质，包括4个主要亚家族，CC趋化因子，CXC趋化因子，CX3C趋化因子和XC趋化因子，这些亚家族中的每一个都有自己的受体：CCRs，CXCRs，CX3CRs和XCRs。

. Our survival study revealed several chemokines and chemokine receptors, including CCL2, CCL3, CCL4, CXCL1 and CXCR4. All of them are risk factors for the prognosis of IPF. Figure

我们的生存研究揭示了几种趋化因子和趋化因子受体，包括CCL2，CCL3，CCL4，CXCL1和CXCR4。所有这些都是IPF预后的危险因素。图

j also revealed a strong interaction between these chemokines and chemokine receptors, suggesting that chemokines play a pivotal role in IPF pathogenesis and might be predictors of the prognosis of IPF, especially CCL2. Figure

j还揭示了这些趋化因子和趋化因子受体之间的强烈相互作用，表明趋化因子在IPF发病机制中起关键作用，可能是IPF预后的预测因子，尤其是CCL2。图

a,b also revealed that the CCL2 expression level in IPF patients is negatively correlated with their lung function (DLco % predicted and FVC % predicted), suggesting that CCL2 expression has a certain predictive value for pulmonary function in patients with IPF.

a、 b还显示IPF患者的CCL2表达水平与肺功能呈负相关（DLco%预测值和FVC%预测值），表明CCL2表达对IPF患者的肺功能具有一定的预测价值。

Alpha-defencins include antimicrobial and cytotoxic peptides that human neutrophils contain in azurophil granules and belong to the mammalian neutrophilic peptide family, and their level has been shown to be elevated in the serum of patients with SSc and is associated with lung involvement

α防御素包括人类嗜中性粒细胞在天青颗粒中含有的抗菌肽和细胞毒性肽，属于哺乳动物嗜中性粒细胞肽家族，其水平在SSc患者血清中升高，并与肺部受累有关

. Our GSEA results revealed enrichment in azurophil granule, and the scRNA-seq results revealed that the CCL2 expression level was upregulated in several cell types including alveolar epithelial type 1 cell, alveolar epithelial type 2 cell, mast cell, ciliated cell, club cell, fibroblast, M1 macrophage, M2 macrophage, monocyte and plasma cell.

我们的GSEA结果显示嗜天青颗粒富集，scRNA-seq结果显示CCL2表达水平在几种细胞类型中上调，包括肺泡上皮1型细胞，肺泡上皮2型细胞，肥大细胞，纤毛细胞，俱乐部细胞，成纤维细胞，M1巨噬细胞，M2巨噬细胞，单核细胞和浆细胞。

These results show that in these cell types of IPF patients, alpha-deficiency-related immune reactions might be activated. Mukae’s study also revealed that plasma alpha-defensin levels were negatively correlated with arterial oxygen tension (PaO2), lung function [vital capacity (%VC), forced expiratory volume in 1 s (FEV1), and carbon monoxide transfer factor (%TLCO)] in IPF patients, indicating that alpha-defencins play important roles in the pathogenesis of IPF.

这些结果表明，在这些细胞类型的IPF患者中，可能会激活与α缺乏相关的免疫反应。Mukae的研究还表明，IPF患者血浆α-防御素水平与动脉血氧分压（PaO2），肺功能[肺活量（%VC），1秒用力呼气量（FEV1）和一氧化碳转移因子（%TLCO）]呈负相关，表明α-防御素在IPF的发病机制中起重要作用。

. We speculate that CCL2 and alpha-defencins interact to mediate the pathogenesis of IPF.

我们推测CCL2和α防御素相互作用以介导IPF的发病机制。

In SSc patients, the GSEA results revealed enrichment of various fatty acid metabolic processes, such as acyl − CoA metabolic process, cholesterol biosynthetic process, fatty − acyl − CoA biosynthetic process, fatty acid synthase activity and fatty acid metabolism, etc. These results indicate that CCL2 may play a role in the pathogenesis of SSc by affecting the metabolism of fatty acids.

在SSc患者中，GSEA结果显示各种脂肪酸代谢过程的富集，例如酰基辅酶A代谢过程，胆固醇生物合成过程，脂肪酰基辅酶A生物合成过程，脂肪酸合酶活性和脂肪酸代谢等。这些结果表明，CCL2可能通过影响脂肪酸的代谢在SSc的发病机制中发挥作用。

Studies have shown that adipose tissue and its secretion group significantly contribute to the pathogenesis of SSc. Adipocytes are directly involved in the pathogenesis of SSc through adipocyte-myofibroblast transformation (AMT) and are the source of myofibroblast precursors at different tissue sites [e.g., dermal white adipose tissue (dWAT) in the skin and lipofibroblasts in the lung].

研究表明，脂肪组织及其分泌组在SSc的发病机制中起着重要作用。脂肪细胞通过脂肪细胞-肌成纤维细胞转化（AMT）直接参与SSc的发病机制，并且是不同组织部位（例如皮肤中的真皮白色脂肪组织（dWAT）和肺中的脂肪成纤维细胞）的肌成纤维细胞前体的来源。

. Moreover, fatty acid metabolism also plays a role in IPF pathogenesis. As mediators of profibrotic reactions, fatty acids and their metabolites (stearic acid, palmitic acid, arachidic acid, etc.) regulate the phenotypic changes in various lung cells during lung tissue maladaptation to remodelling

此外，脂肪酸代谢也在IPF发病机理中起作用。作为促纤维化反应的介质，脂肪酸及其代谢产物（硬脂酸、棕榈酸、花生酸等）在肺组织不适应重塑过程中调节各种肺细胞的表型变化

. Repetitive genetic or environmental stimulation of alveolar epithelial type 2 (AT2) cells can lead to endoplasmic reticulum (ER) stress, which can ultimately drive downstream fibrotic remodelling of IPF lungs, and fatty acid (FA) synthesis and composition are involved in ER stress

. Fatty acid oxidation (FAO) produces a large amount of ATP, provides energy to support the polarization of M2 macrophages, and activates the signalling pathway of macrophage polarization, enabling M2 macrophages to produce profibrotic mediators, such as TGF-β1, activate fibroblasts, and promote extracellular matrix (ECM) deposition.

脂肪酸氧化（FAO）产生大量ATP，提供能量支持M2巨噬细胞的极化，并激活巨噬细胞极化的信号通路，使M2巨噬细胞产生促纤维化介质，如TGF-β1，激活成纤维细胞，促进细胞外基质（ECM）沉积。

. Therefore, the extensive enrichment of fatty acid metabolism in SSc patients with high CCL2 expression suggests possible pulmonary fibrotic injury. In addition, according to our scRNA-seq results, compared with that in healthy donors, CCL2 expression was upregulated in these 3 cell types (alveolar epithelial type 2 cell, M2 macrophage and fibroblast), suggesting that fatty acid metabolism plays a role in the pathogenesis of these 3 cell types in IPF patients..

因此，CCL2高表达的SSc患者脂肪酸代谢的广泛富集提示可能存在肺纤维化损伤。此外，根据我们的scRNA-seq结果，与健康供体相比，这3种细胞类型（肺泡上皮2型细胞，M2巨噬细胞和成纤维细胞）的CCL2表达上调，表明脂肪酸代谢在IPF患者这3种细胞类型的发病机制中起作用。。

One study revealed that the activation of CCL2/CCR2 signalling is involved in the pathogenesis of IPF through the recruitment of macrophages to the lung parenchyma and polarization of the M1 phenotype

一项研究表明，CCL2/CCR2信号的激活通过巨噬细胞向肺实质的募集和M1表型的极化而参与IPF的发病机制

. Our scRNA-seq results revealed that CCL2 expression in M1 and M2 macrophages from patients with IPF was significantly greater than that in healthy donors, suggesting a potential connection between CCL2 and macrophages in the pathogenesis of IPF.

我们的scRNA-seq结果显示，来自IPF患者的M1和M2巨噬细胞中的CCL2表达显着高于健康供体中的CCL2表达，表明CCL2和巨噬细胞在IPF发病机制中可能存在联系。

Our results revealed that CCL2 expression was downregulated in dcSSc patients after 3 months of MMF treatment according to our result. Although this result does not prove a causal relationship between MMF treatment and the downregulation of CCL2 expression, it suggests that MMF may improve SSc by inhibiting the chemokine system, providing a possible direction for the treatment of SSc and even SSc-ILD..

。虽然这一结果并不能证明MMF治疗与CCL2表达下调之间存在因果关系，但它表明MMF可能通过抑制趋化因子系统来改善SSc，为SSc甚至SSc ILD的治疗提供了可能的方向。。

Our study has several limitations. Our study used data from a public database and lacked support from clinical data. Some datasets have small sample sizes, and our study was limited to the transcriptome level, therefore, the results need to be further verified by basic experiments and prospective clinical trials.

我们的研究有几个局限性。我们的研究使用了来自公共数据库的数据，缺乏临床数据的支持。一些数据集的样本量很小，我们的研究仅限于转录组水平，因此，结果需要通过基础实验和前瞻性临床试验进一步验证。

Further studies are needed to elucidate the potential mechanism of CCL2 in the pathogenesis of IPF and SSc..

需要进一步的研究来阐明CCL2在IPF和SSc发病机制中的潜在机制。。

Materials and methods

材料和方法

Data Collection

数据收集

Figure

图

shows the flow chart of the study.

显示了研究的流程图。

The Series Matrix File data file of GSE70866 (platforms: GPL14550 and GPL17077) was downloaded from the Gene Expression Omnibus (GEO) database (

GSE70866的系列矩阵文件数据文件（平台：GPL14550和GPL17077）是从Gene Expression Omnibus（GEO）数据库下载的(

http://www.ncbi.nlm.nih.gov/geo/

), which is from the donors’ bronchoalveolar lavage (BAL) cells and includes 3 cohorts: Freiburg (62 IPF patients and 20 healthy controls), LEUVEN (64 IPF patients), and SIENA (50 IPF patients). The whole GSE70866 expression data was obtained via batch correction and normalization of 3 cohorts via the “removeBatchEffect” function and the “normalizeBetweenArrays” function of the “limma” package in R software 4.3.2.

)，来自供体的支气管肺泡灌洗（BAL）细胞，包括3个队列：弗莱堡（62名IPF患者和20名健康对照），鲁汶（64名IPF患者）和锡耶纳（50名IPF患者）。整个GSE70866表达数据是通过R软件4.3.2中“limma”包的“removeBatchEffect”函数和“normalizeBetweenArrays”函数对3个队列进行批量校正和归一化获得的。

. Before and after batch correction and normalization, PCA was conducted via the PCA function of the “FactoMineR” package version 2.11

。在批量校正和标准化之前和之后，通过“FactoMineR”软件包版本2.11的PCA功能进行PCA

The GSE181549 series matrix file data file (platform: GPL13497), which was obtained from the donors’ skin, was downloaded from GEO. A total of 113 SSc patients and 44 matched healthy controls were included in this dataset; 105 SSc patients had a 2nd biopsy, 76 patients had a 3rd biopsy, and 1 patient had a 4th biopsy.

从捐赠者的皮肤获得的GSE181549系列矩阵文件数据文件（平台：GPL13497）已从GEO下载。该数据集中共包括113名SSc患者和44名匹配的健康对照；105名SSc患者进行了第二次活检，76名患者进行了第三次活检，1名患者进行了第四次活检。

We chose 1st biopsy patients and healthy controls for the following study..

我们选择了第一例活检患者和健康对照进行以下研究。。

5 other datasets of IPF (GSE27957, GSE28042, GSE93606, GSE32537, and GSE122960) and 3 datasets of SSc (GSE181549, GSE58095, and GSE97248) were also downloaded from GEO. The baseline IPF patients’ samples of GSE27957, GSE28042 and GSE93606 was removed batch effect and standardization based on the “removeBatchEffect” function and the “normalizeBetweenArrays” function of “limma” package and of in R software.

还从GEO下载了其他5个IPF数据集（GSE27957，GSE28042，GSE93606，GSE32537和GSE122960）和3个SSc数据集（GSE181549，GSE58095和GSE97248）。基于“limma”软件包和in R软件的“removeBatchEffect”功能和“normalizeBetweenArrays”功能，删除了基线IPF患者的GSE27957，GSE28042和GSE93606样本的批处理效应和标准化。

Table

表

shows the information of all the datasets used in this study.

显示了本研究中使用的所有数据集的信息。

Table 1 The information of all the datasets used in the study.

表1研究中使用的所有数据集的信息。

Full size table

全尺寸表

WGCNA and key module genes identification

WGCNA和关键模块基因鉴定

WGCNA was performed on the GSE70866 and GSE181549 datasets via the “WGCNA” package in R software to screen the gene modules

通过R软件中的“WGCNA”包在GSE70866和GSE181549数据集上进行WGCNA，以筛选基因模块

. First, the median absolute deviation (MAD) of each gene was determined, and 25% of the genes with the smallest MAD and genes whose MAD was less than 0.01 were removed. On the basis of the correlation coefficient R

首先，确定每个基因的中位绝对偏差（MAD），并去除MAD最小的基因和MAD小于0.01的基因的25%。基于相关系数R

> 0.85, the “picksoftthreshold” function was used to construct the gene expression patterns of scale-free networks. After the modules are acquired, according to the module of the first principal component from different module eigengenes (ME), the ME is associated with the clinical features of the evaluation module–trait relationships.

>0.85，“picksoftthreshold”函数用于构建无标度网络的基因表达模式。获得模块后，根据来自不同模块特征基因（ME）的第一主成分的模块，ME与评估模块-性状关系的临床特征相关联。

The modules with the most significant positive and negative correlations of the module-trait relationships were subsequently screened. Finally, the module membership (MM) and gene significance (GS) scores of the modules were also assessed to account for module significance (MS). The first two modules of both datasets were chosen for the following study..

随后筛选模块-性状关系具有最显着正相关和负相关的模块。最后，还评估了模块的模块成员资格（MM）和基因显着性（GS）得分，以说明模块显着性（MS）。两个数据集的前两个模块被选择用于以下研究。。

Construction of a PPI network

PPI网络的构建

The intersecting genes of these 2 groups of modules were then input into the STRING database, and Cytoscape 3.10.0 software was subsequently used to visualize a PPI network. By using the MCC method of the cytoHubba plug-in, the first-ranked 10 genes were chosen for the following study

然后将这两组模块的相交基因输入STRING数据库，随后使用Cytoscape 3.10.0软件可视化PPI网络。通过使用cytoHubba插件的MCC方法，选择排名第一的10个基因进行以下研究

Survival and prognostic analysis

生存和预后分析

Survival information from the GSE70866 dataset was used to analyse the effects of the above 10 genes on the prognosis of patients with IPF. K–M curves of these 10 genes were generated via the “survminer” package (

来自GSE70866数据集的生存信息用于分析上述10个基因对IPF患者预后的影响。这10个基因的K-M曲线是通过“survminer”软件包生成的(

https://cran.r-project.org/web/packages/survminer/index.html

, v.0.4.9) in R software, and the difference in survival between groups was statistically tested via the log‒rank method. Univariate Cox regression of these 10 genes was performed via the “survival” package (

，v.0.4.9），并通过对数秩方法对组间生存率的差异进行统计学检验。(

https://github.com/therneau/survival

, v.3.6–4) and was visualized via the “forestploter” package (

，v.3.6-4），并通过“forestploter”包可视化(

https://github.com/adayim/forestploter

, v.1.1.2) of R software.

。

Machine learning

机器学习

Three machine learning algorithms (LASSO-Cox regression, SVM-RFE and Random Forest) were employed in the baseline IPF samples of GSE70866 to identify potential prognostic biomarkers from a set of 10 genes identified in the PPI network. LASSO regression, SVM-RFE and random forest algorithms were run on SSc and control samples from GSE53845 to screen for potential SSc diagnostic biomarkers.

在GSE70866的基线IPF样本中使用了三种机器学习算法（LASSO-Cox回归，SVM-RFE和随机森林），以从PPI网络中鉴定的一组10个基因中鉴定潜在的预后生物标志物。LASSO回归，SVM-RFE和随机森林算法在来自GSE53845的SSc和对照样品上运行，以筛选潜在的SSc诊断生物标志物。

LASSO-Cox and LASSO regression can improve prediction accuracy and minimize the risk of overfitting by selecting variables.

LASSO-Cox和LASSO回归可以通过选择变量来提高预测准确性并最小化过度拟合的风险。

. The random forest model is an effective prediction tool with high accuracy, sensitivity and specificity, is independent of variable conditions, and can help assess the importance of variables

。随机森林模型是一种有效的预测工具，具有较高的准确性，敏感性和特异性，与可变条件无关，可以帮助评估变量的重要性

. SVM-RFE enables variable selection and interpretation of associations between predictive and response variables when analysing biomedical data and is achieved with high levels of accuracy and speed and low computational costs

在分析生物医学数据时，SVM-RFE可以对预测变量和响应变量之间的关联进行变量选择和解释，并且具有较高的准确性和速度以及较低的计算成本

. These 3 algorithms were conducted according to the R packages “glmnet”, “randomForest” and “e1071” (

。这3种算法是根据R包“glmnet”，“randomForest”和“e1071”进行的(

https://github.com/cran/e1071)

. The genes common to these algorithms are considered potential central genes for the prognosis of IPF and the diagnosis of SSc.

这些算法共有的基因被认为是IPF预后和SSc诊断的潜在中心基因。

Single-cell RNA data processing and clustering

单细胞RNA数据处理和聚类

The raw data of GSE122960 was downloaded from the GEO database for single-cell sequencing analysis. 4 IPF samples (GSM3489183, GSM3489184, GSM3489188, and GSM3489190) and 4 healthy samples (GSM3489182, GSM3489185, GSM3489189, and GSM3489191) were chosen according to the donors’ ages

GSE122960的原始数据是从GEO数据库下载的，用于单细胞测序分析。根据捐赠者的年龄选择了4个IPF样本（GSM3489183、GSM3489184、GSM3489188和GSM3489190）和4个健康样本（GSM3489182、GSM3489185、GSM3489189和GSM34891911）

. The 8 samples of GSE122960 were analysed according to the “Seurat” package version 5.1.0

。根据“Seurat”软件包版本5.1.0分析了8个GSE122960样品

. The cells were filtered with the criteria of 200 < number of feature RNA < 5000, percent of mt < 20, number of genes in single cell < 30,000, percentage of ribosomal genes > 3 and percentage of hemoglobin genes < 0.1. The FindVariableFeatures function was used, and the top 2000 hypervariable genes were selected on the basis of the variance stabilization transformation (vst) method.

.用200个特征RNA 5000的标准过滤细胞，mt 20的百分比，单细胞基因数<30000，核糖体基因百分比>3和血红蛋白基因百分比 < 0.1. 使用FindVariableFeatures函数，并基于方差稳定变换（vst）方法选择前2000个高变基因。

The top 2000 genes selected previously were scaled by using the ScaleData function, and the RunPCA function was subsequently used to reduce the dimension of the PCA. Dim = 1:15 was chosen for the following analysis. The IntegrateLayers function and HarmonyIntegration method in the “Seurat” package version 5.1.0.

先前选择的前2000个基因通过使用ScaleData函数进行缩放，随后使用RunPCA函数来减小PCA的维数。选择Dim=1:15进行以下分析。“Seurat”包版本5.1.0中的IntegrateLayers函数和HarmonyIntegration方法。

were subsequently used to remove the batch effect between samples.

随后用于消除样品之间的批次效应。

. The cells were then clustered into 27 clusters via the FindNeighbors and FindClusters functions with a resolution = 0.8. We used the RunUMAP function for visualization, and the cell type annotation was based on the CellMarker2.0 database (

然后通过FindNeighbors和FindClusters函数将细胞聚类为27个簇，并具有分辨率 = 0.8. 我们使用RunUMAP函数进行可视化，细胞类型注释基于CellMarker2.0数据库(

http://117.50.127.228/CellMarker/

), Sun et al.’s study and the original article of GSE122960

)，Sun等人的研究和GSE122960的原始文章

Enrichment analysis

富集分析

GSEA is a reliable enrichment method based on gene expression levels

GSEA是一种基于基因表达水平的可靠富集方法

. In our research, we calculated the cut-off level of CCL2 via the “survival” package on the basis of the gene expression level and prognostic information of the merged dataset of GSE27957, GSE28042, and GSE93606, and divided the IPF samples into 2 groups on the basis of the cut-off level. Additionally, the SSc samples in GSE58095 were divided into 2 groups on the basis of the median CCL2 expression level.

在我们的研究中，我们根据GSE27957，GSE28042和GSE93606合并数据集的基因表达水平和预后信息，通过“生存”包计算了CCL2的临界水平，并根据临界水平将IPF样本分为2组。此外，根据中位CCL2表达水平，将GSE58095中的SSc样品分为2组。

We used the “ClusterProfiler” package for GO (Gene Ontology) and KEGG (Kyoto Encyclopedia of Genes and Genomes) analysis via GSEA enrichment analysis of these 2 datasets, and p values < 0.05 and FDRs < 0.25 were considered statistically significant.

我们通过对这两个数据集的GSEA富集分析，使用“ClusterProfiler”软件包进行GO（基因本体论）和KEGG（京都基因和基因组百科全书）分析，p值<0.05和FDRs<0.25被认为具有统计学意义。

Statistical analysis

统计分析

The analyses and visualization in this research were performed in R software (version 4.3.2) and GraphPad Prism 9.5.0. The linear fit was performed via the Levenberg‒Marquardt method. Based on whether the data follows a normal distribution and whether it is a paired design, we used a t test or Mann‒Whitney U test or Wilcoxon matched-pairs signed-rank test to determine significant differences between the groups, and .

本研究中的分析和可视化是在R软件（版本4.3.2）和GraphPad Prism 9.5.0中进行的。通过Levenberg-Marquardt方法进行线性拟合。基于数据是否遵循正态分布以及是否是配对设计，我们使用t检验或Mann-Whitney U检验或Wilcoxon配对符号秩检验来确定各组之间的显着差异。

< 0.05 was the criterion for statistical significance.

。

Conclusion

结论

Our study revealed the common characteristic gene CCL2 in IPF and SSc and its expression was negatively correlated with the survival and lung function of IPF patients. In addition, CCL2 might be a potential target of MMF in the treatment of dcSSc. The mechanism of CCL2 in the pathogenesis of IPF and SSc may be related to TGF-β pathway, chemokine system, alpha-defencins and fatty acid metabolic processes.

我们的研究揭示了IPF和SSc中的共同特征基因CCL2，其表达与IPF患者的生存和肺功能呈负相关。此外，CCL2可能是MMF治疗dcSSc的潜在靶点。CCL2在IPF和SSc发病机制中的作用可能与TGF-β途径、趋化因子系统、α防御素和脂肪酸代谢过程有关。

Fibroblasts, monocytes, M1 macrophages and M2 macrophages may play a role in the pathogenesis of CCL2 in relation to IPF. Our study provides a possible direction for the pathogenesis and treatment of SSc and IPF..

成纤维细胞，单核细胞，M1巨噬细胞和M2巨噬细胞可能在CCL2与IPF相关的发病机制中起作用。。。

Data availability

数据可用性

The datasets extracted and/or analysed during the current study are available in the GEO repository, (

在当前研究期间提取和/或分析的数据集可在GEO存储库中找到(

https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE70866

https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE27957

https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE28042

https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE93606

https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE32537

https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE122960

https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE181549

https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE58095

https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE97248

References

参考文献

Raghu, G. et al. An official ATS/ERS/JRS/ALAT statement: Idiopathic pulmonary fibrosis: Evidence-based guidelines for diagnosis and management.

Raghu，G。等。ATS/ERS/JRS/ALAT官方声明：特发性肺纤维化：基于证据的诊断和管理指南。

Am. J. Respir. Crit. Care Med.

我是J.呼吸。Crit。护理医学。

183

, 788–824.

https://doi.org/10.1164/rccm.2009-040GL

(2011).

Article

文章

PubMed

PubMed Central

公共医学中心

Google Scholar

谷歌学者

Richeldi, L., Rubin, A. S., Avdeev, S., Udwadia, Z. F. & Xu, Z. J. Idiopathic pulmonary fibrosis in BRIC countries: the cases of Brazil, Russia, India, and China.

Richeldi，L.，Rubin，A.S.，Avdeev，S.，Udwadia，Z.F。＆Xu，Z.J。金砖四国的特发性肺纤维化：巴西，俄罗斯，印度和中国的病例。

BMC Med.

BMC医学。

, 237.

https://doi.org/10.1186/s12916-015-0495-0

(2015).

Article

文章

PubMed

PubMed Central

公共医学中心

Google Scholar

谷歌学者

Raghu, G. et al. An Official ATS/ERS/JRS/ALAT clinical practice guideline: Treatment of idiopathic pulmonary fibrosis an update of the 2011 clinical practice guideline.

Raghu，G.等人，《官方ATS/ERS/JRS/ALAT临床实践指南：特发性肺纤维化的治疗——2011年临床实践指南的更新》。

Am. J. Respir. Crit. Care Med.

我是J.呼吸。Crit。护理医学。

192

, 13–119.

https://doi.org/10.1164/rccm.201506-1063ST

(2015).

Article

文章

MATH

数学

Google Scholar

谷歌学者

Martinez, F. J. et al. Idiopathic pulmonary fibrosis.

Martinez，F.J.等人。特发性肺纤维化。

Nat. Rev. Dis. Primers

自然版本Dis。引物

, 17074.

https://doi.org/10.1038/nrdp.2017.74

(2017).

Article

文章

PubMed

Google Scholar

谷歌学者

Hewlett, J. C., Kropski, J. A. & Blackwell, T. S. Idiopathic pulmonary fibrosis: Epithelial-mesenchymal interactions and emerging therapeutic targets.

Hewlett，J.C.，Kropski，J.A。＆Blackwell，T.S。特发性肺纤维化：上皮-间质相互作用和新兴治疗靶点。

Matrix Biol. J. Int. Soc. Matrix Biol.

矩阵生物学。J、国际社会矩阵生物学。

71–72

, 112–127.

https://doi.org/10.1016/j.matbio.2018.03.021

(2018).

Article

文章

CAS

中科院

Google Scholar

谷歌学者

Liu, G. et al. Therapeutic targets in lung tissue remodelling and fibrosis.

Liu，G.等人。肺组织重塑和纤维化的治疗靶点。

Pharmacol. Therapeut.

药理学。治疗专家。

225

, 107839.

https://doi.org/10.1016/j.pharmthera.2021.107839

(2021).

Article

文章

CAS

中科院

MATH

数学

Google Scholar

谷歌学者

Farkas, L. et al. VEGF ameliorates pulmonary hypertension through inhibition of endothelial apoptosis in experimental lung fibrosis in rats.

Farkas，L。等人。VEGF通过抑制大鼠实验性肺纤维化中的内皮细胞凋亡来改善肺动脉高压。

J. Clin. Investig.

J、克林。投资。

119

, 1298–1311.

https://doi.org/10.1172/jci36136

(2009).

Article

文章

CAS

中科院

PubMed

PubMed Central

公共医学中心

MATH

数学

Google Scholar

谷歌学者

Elhai, M. et al. Mapping and predicting mortality from systemic sclerosis.

Elhai，M.等人。绘制和预测系统性硬化症的死亡率。

Ann. Rheumatic Dis.

风湿性疾病。

, 1897–1905.

https://doi.org/10.1136/annrheumdis-2017-211448

(2017).

Article

文章

MATH

数学

Google Scholar

谷歌学者

Volkmann, E. R., Andréasson, K. & Smith, V. Systemic sclerosis.

Volkmann，E.R.，Andréasson，K。＆Smith，V。系统性硬化症。

Lancet (London, England)

柳叶刀（英国伦敦）

401

, 304–318.

https://doi.org/10.1016/s0140-6736(22)01692-0

(2023).

Article

文章

PubMed

Google Scholar

谷歌学者

Jerjen, R., Nikpour, M., Krieg, T., Denton, C. P. & Saracino, A. M. Systemic sclerosis in adults. Part I: Clinical features and pathogenesis.

Jerjen，R.，Nikpour，M.，Krieg，T.，Denton，C.P。＆Saracino，A.M。成人系统性硬化症。。

J. Am. Acad. Dermatol.

, 937–954.

https://doi.org/10.1016/j.jaad.2021.10.065

(2022).

Article

文章

PubMed

Google Scholar

谷歌学者

Airò, P. et al. Prevalence of systemic sclerosis in Valtrompia in northern Italy. A collaborative study of rheumatologists and general practitioners.

Airò，P.等人。意大利北部Valtrompia的系统性硬化症患病率。风湿病学家和全科医生的合作研究。

Clin. Exp. Rheumatol.

临床。实验风湿病。

, 878–880 (2007).

PubMed

MATH

数学

Google Scholar

谷歌学者

Hoffmann-Vold, A. M., Midtvedt, Ø., Molberg, Ø., Garen, T. & Gran, J. T. Prevalence of systemic sclerosis in south-east Norway.

霍夫曼·沃尔德，上午，米德维特，Ø。，莫尔伯格。，Garen，T。＆Gran，J.T。挪威东南部系统性硬化症的患病率。

Rheumatology (Oxford, England)

风湿病学（英国牛津）

, 1600–1605.

https://doi.org/10.1093/rheumatology/kes076

(2012).

Article

文章

PubMed

Google Scholar

谷歌学者

Volkmann, E. R., Andréasson, K. & Smith, V. Systemic sclerosis.

Volkmann，E.R.，Andréasson，K。＆Smith，V。系统性硬化症。

Lancet

柳叶刀

401

, 304–318.

https://doi.org/10.1016/s0140-6736(22)01692-0

(2023).

Article

文章

PubMed

Google Scholar

谷歌学者

Lomelí-Nieto, J. A. et al. Transforming growth factor beta isoforms and TGF-βR1 and TGF-βR2 expression in systemic sclerosis patients.

Lomelí-Nieto，J.A.等人。系统性硬化症患者的转化生长因子β亚型和TGF-βR1和TGF-βR2表达。

Clin. Exp. Med.

克林顿。实验医学。

, 471–481.

https://doi.org/10.1007/s10238-022-00841-0

(2022).

Article

文章

CAS

中科院

PubMed

Google Scholar

谷歌学者

Tabata, K. et al. Up-regulation of IGF-1, RANTES and VEGF in patients with anti-centromere antibody-positive early/mild systemic sclerosis.

Tabata，K。等人。抗着丝粒抗体阳性早期/轻度系统性硬化症患者IGF-1，RANTES和VEGF的上调。

Mod. Rheumatol.

勇气。风湿病。

, 171–176.

https://doi.org/10.1080/14397595.2020.1726599

(2021).

Article

文章

CAS

中科院

PubMed

MATH

数学

Google Scholar

谷歌学者

Mao, J. et al. Hypoxia-induced interstitial transformation of microvascular endothelial cells by mediating HIF-1Î±/VEGF signaling in systemic sclerosis.

Mao，J。等人。缺氧通过介导系统性硬化症中的HIF-1Î±/VEGF信号传导诱导微血管内皮细胞的间质转化。

PLoS ONE

公共科学图书馆一号

, e0263369.

，e0263369。

https://doi.org/10.1371/journal.pone.0263369

(2022).

Article

文章

CAS

中科院

PubMed

PubMed Central

公共医学中心

Google Scholar

谷歌学者

Pope, J. E. et al. State-of-the-art evidence in the treatment of systemic sclerosis.

Pope，J.E.等人。系统性硬化症治疗的最新证据。

Nat. Rev. Rheumatol.

风湿病杂志。

, 212–226.

https://doi.org/10.1038/s41584-023-00909-5

(2023).

Article

文章

CAS

中科院

PubMed

PubMed Central

公共医学中心

MATH

数学

Google Scholar

谷歌学者

Fernández-Codina, A., Walker, K. M. & Pope, J. E. Treatment Algorithms for Systemic Sclerosis According to Experts.

Fernández-Codina，A.，Walker，K.M。和Pope，J.E。根据专家的说法，系统性硬化症的治疗算法。

Arthritis Rheumatol. (Hoboken, N.J.)

风湿性关节炎。（新泽西州霍博肯）

, 1820–1828.

https://doi.org/10.1002/art.40560

(2018).

Article

文章

CAS

中科院

MATH

数学

Google Scholar

谷歌学者

Nie, Y. et al. NFATc3 Promotes Pulmonary Inflammation and Fibrosis by Regulating Production of CCL2 and CXCL2 in Macrophages.

Nie，Y。等人。NFATc3通过调节巨噬细胞中CCL2和CXCL2的产生来促进肺部炎症和纤维化。

Aging Dis.

老化Dis。

, 1441–1457.

https://doi.org/10.14336/ad.2022.1202

(2023).

Article

文章

PubMed

PubMed Central

公共医学中心

MATH

数学

Google Scholar

谷歌学者

Bian, F. et al. Lung endothelial cells regulate pulmonary fibrosis through FOXF1/R-Ras signaling.

。

Nat. Commun.

Nat.普通。

, 2560.

https://doi.org/10.1038/s41467-023-38177-2

(2023).

Article

文章

ADS

CAS

中科院

PubMed

PubMed Central

公共医学中心

Google Scholar

谷歌学者

Yuan, X. et al. Human hypofunctional NCF1 variants promote pulmonary fibrosis in the bleomycin-induced mouse model and patients with systemic sclerosis via expansion of SPP1(+) monocytes-derived macrophages.

Yuan，X。等人。人类功能低下的NCF1变体通过扩增SPP1（+）单核细胞衍生的巨噬细胞促进博来霉素诱导的小鼠模型和系统性硬化症患者的肺纤维化。

Ann. Rheumatic Dis.

风湿性疾病。

https://doi.org/10.1136/ard-2024-226034

(2024).

Article

文章

Google Scholar

谷歌学者

Van Coillie, E., Van Damme, J. & Opdenakker, G. The MCP/eotaxin subfamily of CC chemokines.

Van Coillie，E.，Van Damme，J。＆Opdenakker，G。CC趋化因子的MCP/eotaxin亚家族。

Cytokine Growth Factor Rev.

细胞因子生长因子Rev。

, 61–86.

https://doi.org/10.1016/s1359-6101(99)00005-2

(1999).

Article

文章

PubMed

Google Scholar

谷歌学者

Miyabe, Y., Lian, J., Miyabe, C. & Luster, A. D. Chemokines in rheumatic diseases: pathogenic role and therapeutic implications.

Miyabe，Y.，Lian，J.，Miyabe，C。＆Luster，A.D。趋化因子在风湿性疾病中的致病作用和治疗意义。

Nat. Rev. Rheumatol.

风湿病杂志。

, 731–746.

https://doi.org/10.1038/s41584-019-0323-6

(2019).

Article

文章

PubMed

MATH

数学

Google Scholar

谷歌学者

Rockey, D. C., Bell, P. D. & Hill, J. A. Fibrosis–a common pathway to organ injury and failure.

Rockey，D.C.，Bell，P.D。和Hill，J.A。纤维化-器官损伤和衰竭的常见途径。

New Engl. J. Med.

新英语。医学杂志。

372

, 1138–1149.

https://doi.org/10.1056/NEJMra1300575

(2015).

Article

文章

CAS

中科院

PubMed

Google Scholar

谷歌学者

Murray, L. A. et al. Hyper-responsiveness of IPF/UIP fibroblasts: Interplay between TGFβ1, IL-13 and CCL2.

Murray，L.A。等人。IPF/UIP成纤维细胞的高反应性：TGFβ1，IL-13和CCL2之间的相互作用。

Int. J. Biochem. Cell Biol.

Int.J.生物化学。细胞生物学。

, 2174–2182.

https://doi.org/10.1016/j.biocel.2008.02.016

(2008).

Article

文章

CAS

中科院

PubMed

MATH

数学

Google Scholar

谷歌学者

Laurent, P. et al. Interleukin-1β-activated microvascular endothelial cells promote DC-SIGN-positive alternatively activated macrophages as a mechanism of skin fibrosis in systemic sclerosis.

Laurent，P。等人。白细胞介素-1β激活的微血管内皮细胞促进DC-SIGN阳性交替激活的巨噬细胞作为系统性硬化症皮肤纤维化的机制。

Arthritis Rheumatol. (Hoboken, N.J.)

风湿性关节炎。（新泽西州霍博肯）

, 1013–1026.

https://doi.org/10.1002/art.42061

(2022).

Article

文章

CAS

中科院

Google Scholar

谷歌学者

Bandinelli, F. et al. CCL2, CCL3 and CCL5 chemokines in systemic sclerosis: the correlation with SSc clinical features and the effect of prostaglandin E1 treatment.

Bandinelli，F。等。系统性硬化症中的CCL2，CCL3和CCL5趋化因子：与SSc临床特征的相关性以及前列腺素E1治疗的效果。

Clin. Exp. Rheumatol.

临床。实验风湿病。

, S44-49 (2012).

，S44-49（2012）。

PubMed

MATH

数学

Google Scholar

谷歌学者

Kass, D. J. et al. Lung gene expression and single cell analyses reveal two subsets of idiopathic pulmonary fibrosis (IPF) patients associated with different pathogenic mechanisms.

Kass，D.J.等人，肺基因表达和单细胞分析揭示了与不同致病机制相关的特发性肺纤维化（IPF）患者的两个子集。

PLoS ONE

公共科学图书馆一号

, e0248889.

，e0248889。

https://doi.org/10.1371/journal.pone.0248889

(2021).

Article

文章

CAS

中科院

Google Scholar

谷歌学者

Hughes, C. E. & Nibbs, R. J. B. A guide to chemokines and their receptors.

Hughes，C.E。＆Nibbs，R.J.B。趋化因子及其受体指南。

FEBS J.

FEBS J。

285

, 2944–2971.

https://doi.org/10.1111/febs.14466

(2018).

Article

文章

CAS

中科院

PubMed

PubMed Central

公共医学中心

MATH

数学

Google Scholar

谷歌学者

Mukae, H. et al. Raised plasma concentrations of alpha-defensins in patients with idiopathic pulmonary fibrosis.

Mukae，H.等人提高了特发性肺纤维化患者的血浆α-防御素浓度。

Thorax

胸部

, 623–628.

https://doi.org/10.1136/thorax.57.7.623

(2002).

Article

文章

CAS

中科院

PubMed

PubMed Central

公共医学中心

MATH

数学

Google Scholar

谷歌学者

Gedik, T. E. et al. Serum defensin levels in patients with systemic sclerosis.

Gedik，T.E.等人。系统性硬化症患者的血清防御素水平。

Adv. Rheumatol. (London, England)

风湿病顾问。（英国伦敦）

, 54.

https://doi.org/10.1186/s42358-020-00156-2

(2020).

Article

文章

MATH

数学

Google Scholar

谷歌学者

Brezovec, N., Burja, B. & Lakota, K. Adipose tissue and adipose secretome in systemic sclerosis.

Brezovec，N.，Burja，B。＆Lakota，K。系统性硬化症中的脂肪组织和脂肪分泌组。

Curr. Opin. Rheumatol.

。奥平。风湿病。

, 505–513.

https://doi.org/10.1097/bor.0000000000000838

(2021).

Article

文章

CAS

中科院

PubMed

Google Scholar

谷歌学者

Wygrecka, M. et al. Decoding the role of fatty acids and their metabolites in lung fibrosis.

Wygrecka，M.等人。解码脂肪酸及其代谢物在肺纤维化中的作用。

Pol. Arch. Internal Med.

波尔。拱门。内部医学。

133

, 16520.

https://doi.org/10.20452/pamw.16520

(2023).

Article

文章

MATH

数学

Google Scholar

谷歌学者

Han, J. & Kaufman, R. J. The role of ER stress in lipid metabolism and lipotoxicity.

Han，J。＆Kaufman，R.J。内质网应激在脂质代谢和脂毒性中的作用。

J. Lipid Res.

J、脂质研究。

, 1329–1338.

https://doi.org/10.1194/jlr.R067595

(2016).

Article

文章

CAS

中科院

PubMed

PubMed Central

公共医学中心

MATH

数学

Google Scholar

谷歌学者

Katzen, J. & Beers, M. F. Contributions of alveolar epithelial cell quality control to pulmonary fibrosis.

Katzen，J。＆Beers，M.F。肺泡上皮细胞质量控制对肺纤维化的贡献。

J. Clin. Investig.

J、克林。投资。

130

, 5088–5099.

https://doi.org/10.1172/jci139519

(2020).

Article

文章

CAS

中科院

PubMed

PubMed Central

公共医学中心

MATH

数学

Google Scholar

谷歌学者

Geng, J., Liu, Y., Dai, H. & Wang, C. Fatty Acid Metabolism and Idiopathic Pulmonary Fibrosis.

耿，J.，刘，Y.，戴，H。＆Wang，C。脂肪酸代谢和特发性肺纤维化。

Front. Physiol.

正面。生理学。

, 794629.

https://doi.org/10.3389/fphys.2021.794629

(2022).

Article

文章

PubMed

PubMed Central

公共医学中心

Google Scholar

谷歌学者

Li, C. et al. Tetrahedral DNA loaded siCCR2 restrains M1 macrophage polarization to ameliorate pulmonary fibrosis in chemoradiation-induced murine model.

Li，C。等人。四面体DNA负载的siCCR2抑制M1巨噬细胞极化，以改善放化疗诱导的小鼠模型中的肺纤维化。

Mol. Ther.

摩尔热。

, 766–782.

https://doi.org/10.1016/j.ymthe.2024.01.022

(2024).

Article

文章

CAS

中科院

PubMed

Google Scholar

谷歌学者

Ritchie, M. E. et al. limma powers differential expression analyses for RNA-sequencing and microarray studies.

Ritchie，M.E.等人limma为RNA测序和微阵列研究提供了差异表达分析的能力。

Nucleic Acids Res.

核酸研究。

, e47.

，e47。

https://doi.org/10.1093/nar/gkv007

(2015).

Article

文章

CAS

中科院

PubMed

PubMed Central

公共医学中心

MATH

数学

Google Scholar

谷歌学者

Lê, S., Josse, J. & Husson, F. FactoMineR: An R Package for Multivariate Analysis.

Lê，S.，Josse，J。＆Husson，F。FactoMineR：用于多变量分析的R包。

J. Stat. Softw.

J.统计软件。

, 1–18.

https://doi.org/10.18637/jss.v025.i01

(2008).

Article

文章

MATH

数学

Google Scholar

谷歌学者

Langfelder, P. & Horvath, S. WGCNA: An R package for weighted correlation network analysis.

Langfelder，P。＆Horvath，S。WGCNA：用于加权相关网络分析的R包。

BMC Bioinform.

BMC生物信息。

, 559.

https://doi.org/10.1186/1471-2105-9-559

(2008).

Article

文章

CAS

中科院

MATH

数学

Google Scholar

谷歌学者

Chin, C. H. et al. cytoHubba: identifying hub objects and sub-networks from complex interactome.

Chin，C.H.等人。cytoHubba：从复杂的相互作用组中识别中枢对象和子网络。

BMC Syst. Biol.

BMC系统。生物。

8 Suppl 4

8供应4

, S11.

，S11。

https://doi.org/10.1186/1752-0509-8-s4-s11

(2014).

Article

文章

ADS

PubMed

MATH

数学

Google Scholar

谷歌学者

Tibshirani, R. The lasso method for variable selection in the Cox model.

。

Stat. Med.

统计医学。

, 385–395.

https://doi.org/10.1002/(sici)1097-0258(19970228)16:4%3c385::aid-sim380%3e3.0.co;2-3

https://doi.org/10.1002/（sici）1097-0258（1997028）16:4%3c385：：aid-sim380%3E3.0.co；2-3

(1997).

Article

文章

CAS

中科院

PubMed

MATH

数学

Google Scholar

谷歌学者

Ellis, K. et al. A random forest classifier for the prediction of energy expenditure and type of physical activity from wrist and hip accelerometers.

Ellis，K.等人。一种随机森林分类器，用于预测腕部和臀部加速度计的能量消耗和身体活动类型。

Physiol. Meas.

。米亚斯。

, 2191–2203.

https://doi.org/10.1088/0967-3334/35/11/2191

(2014).

Article

文章

PubMed

PubMed Central

公共医学中心

MATH

数学

Google Scholar

谷歌学者

Nakao, H. et al. Determination of individual factors associated with hallux valgus using SVM-RFE.

Nakao，H.等人。使用SVM-RFE确定与拇外翻相关的个体因素。

BMC Musculoskeletal Disord.

BMC肌肉骨骼疾病。

, 534.

https://doi.org/10.1186/s12891-023-06303-2

(2023).

Article

文章

MATH

数学

Google Scholar

谷歌学者

Friedman, J., Hastie, T. & Tibshirani, R. Regularization paths for generalized linear models via coordinate descent.

Friedman，J.，Hastie，T。＆Tibshirani，R。通过坐标下降的广义线性模型的正则化路径。

J. Stat. Softw.

J.统计软件。

, 1–22 (2010).

Article

文章

PubMed

PubMed Central

公共医学中心

MATH

数学

Google Scholar

谷歌学者

Liaw, A. & Wiener, M. C. Classification and regression by randomForest.

Liaw，A。＆Wiener，M.C。通过随机森林进行分类和回归。

R News.

R新闻。

, 18–22 (2022).

Reyfman, P. A. et al. Single-cell transcriptomic analysis of human lung provides insights into the pathobiology of pulmonary fibrosis.

Reyfman，P.A。等人。人肺的单细胞转录组学分析提供了对肺纤维化病理生物学的见解。

Am. J. Respir. Crit. Care Med.

我是J.呼吸。Crit。护理医学。

199

, 1517–1536.

https://doi.org/10.1164/rccm.201712-2410OC

(2019).

Article

文章

CAS

中科院

PubMed

PubMed Central

公共医学中心

MATH

数学

Google Scholar

谷歌学者

Hao, Y. et al. Dictionary learning for integrative, multimodal and scalable single-cell analysis.

Hao，Y。等人。用于综合，多模式和可扩展单细胞分析的词典学习。

Nat. Biotechnol.

国家生物技术。

, 293–304.

https://doi.org/10.1038/s41587-023-01767-y

(2024).

Article

文章

CAS

中科院

PubMed

MATH

数学

Google Scholar

谷歌学者

Sun, X. et al. A census of the lung: Cell Cards from LungMAP.

。

Dev Cell

开发单元

, 112-145.e112.

，112-145.e112。

https://doi.org/10.1016/j.devcel.2021.11.007

(2022).

Article

文章

CAS

中科院

PubMed

MATH

数学

Google Scholar

谷歌学者

Subramanian, A. et al. Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles.

Subramanian，A。等人。基因集富集分析：一种基于知识的方法，用于解释全基因组表达谱。

Proc. Natl. Acad. Sci. USA

Proc。纳特尔。阿卡德。滑雪。美国

102

, 15545–15550.

https://doi.org/10.1073/pnas.0506580102

(2005).

Article

文章

ADS

CAS

中科院

PubMed

PubMed Central

公共医学中心

Google Scholar

谷歌学者

Yu, G., Wang, L. G., Han, Y. & He, Q. Y. clusterProfiler: an R package for comparing biological themes among gene clusters.

Yu，G.，Wang，L.G.，Han，Y。＆He，Q.Y。clusterProfiler：用于比较基因簇之间生物学主题的R包。

Omics A J. Integr. Biol.

组学A J.Integr。生物。

, 284–287.

https://doi.org/10.1089/omi.2011.0118

(2012).

Article

文章

CAS

中科院

MATH

数学

Google Scholar

谷歌学者

Kanehisa, M. & Goto, S. KEGG: kyoto encyclopedia of genes and genomes.

Kanehisa，M。＆Goto，S。KEGG：京都基因与基因组百科全书。

Nucleic Acids Res.

核酸研究。

, 27–30.

https://doi.org/10.1093/nar/28.1.27

(2000).

Article

文章

CAS

中科院

PubMed

PubMed Central

公共医学中心

MATH

数学

Google Scholar

谷歌学者

Kanehisa, M. Toward understanding the origin and evolution of cellular organisms.

Kanehisa，M。了解细胞生物的起源和进化。

Protein Sci. Publ. Protein Soc.

蛋白质科学。公共。蛋白质Soc。

, 1947–1951.

https://doi.org/10.1002/pro.3715

(2019).

Article

文章

CAS

中科院

MATH

数学

Google Scholar

谷歌学者

Kanehisa, M., Furumichi, M., Sato, Y., Kawashima, M. & Ishiguro-Watanabe, M. KEGG for taxonomy-based analysis of pathways and genomes.

。

Nucleic Acids Res.

核酸研究。

, D587-d592.

，D587-d592。

https://doi.org/10.1093/nar/gkac963

(2023).

Article

文章

CAS

中科院

PubMed

Google Scholar

谷歌学者

Download references

下载参考资料

Funding

资金

This research was supported in part by 2022 Heilongjiang Province key research and development plan project (JD22C008).

这项研究得到了2022年黑龙江省重点研究发展计划项目（JD22C008）的部分支持。

Author information

作者信息

Author notes

作者注释

These authors contributed equally: Ning Shan, Yu Shang and Yaowu He.

这些作者做出了同样的贡献：宁山，余尚和姚武河。

Authors and Affiliations

作者和隶属关系

Harbin Medical University, Harbin, Heilongjiang Province, China

哈尔滨医科大学，黑龙江省哈尔滨

Ning Shan, Yaowu He, Zhe Wen, Shangwei Ning & Hong Chen

宁山、何耀武、哲文、宁尚伟和陈红

Department of Pulmonary and Critical Care Medicine, The Second Affiliated Hospital of Harbin Medical University, Harbin, Heilongjiang Province, China

哈尔滨医科大学第二附属医院肺与重症医学科，黑龙江省哈尔滨

Ning Shan, Yaowu He, Zhe Wen & Hong Chen

宁山、何耀武、浙文和陈红

The Second Hospital of Heilongjiang Province, Harbin, Heilongjiang Province, China

黑龙江省第二医院，黑龙江省哈尔滨

Yu Shang

余尚（音）

College of Bioinformatics Science and Technology, Harbin Medical University, Harbin, Heilongjiang Province, China

哈尔滨医科大学生物信息科学与技术学院，黑龙江省哈尔滨

Shangwei Ning

宁尚伟

Authors

作者

Ning Shan

宁山

View author publications

查看作者出版物

You can also search for this author in

您也可以在中搜索此作者

PubMed

Google Scholar

谷歌学者

Yu Shang

余尚（音）

View author publications

查看作者出版物

You can also search for this author in

您也可以在中搜索此作者

PubMed

Google Scholar

谷歌学者

Yaowu He

何耀武

View author publications

查看作者出版物

You can also search for this author in

您也可以在中搜索此作者

PubMed

Google Scholar

谷歌学者

Zhe Wen

哲文

View author publications

查看作者出版物

You can also search for this author in

您也可以在中搜索此作者

PubMed

Google Scholar

谷歌学者

Shangwei Ning

宁尚伟

View author publications

查看作者出版物

You can also search for this author in

您也可以在中搜索此作者

PubMed

Google Scholar

谷歌学者

Hong Chen

陈红（音）

View author publications

查看作者出版物

You can also search for this author in

您也可以在中搜索此作者

PubMed

Google Scholar

谷歌学者

Contributions

捐款

N.S. wrote the main manuscript text, performed the data analysis, Y.S. and Y.H. determined the research process, reviewed and edited the manuscript, Z.W. performed the visualization, HC provided financial support, and S.N. administered this project. All authors contributed to the article and approved the submitted version..

N、 S.撰写了主要手稿文本，进行了数据分析，Y.S.和Y.H.确定了研究过程，审查和编辑了手稿，Z.W.进行了可视化，HC提供了财务支持，S.N.管理了这个项目。所有作者都为这篇文章做出了贡献，并批准了提交的版本。。

Corresponding authors

通讯作者

Correspondence to

通信对象

Shangwei Ning

宁尚伟

或

Hong Chen

陈红（音）

Ethics declarations

道德宣言

Competing interests

相互竞争的利益

The authors declare no competing interests.

作者声明没有利益冲突。

Additional information

其他信息

Publisher’s note

出版商注释

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

Springer Nature在已发布的地图和机构隶属关系中的管辖权主张方面保持中立。

Rights and permissions

权限和权限

Open Access

开放存取

This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.

本文根据知识共享署名4.0国际许可证进行许可，该许可证允许以任何媒体或格式使用，共享，改编，分发和复制，只要您对原始作者和来源给予适当的信任，提供知识共享许可证的链接，并指出是否进行了更改。

The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

To view a copy of this licence, visit .

要查看此许可证的副本，请访问。

http://creativecommons.org/licenses/by/4.0/

Reprints and permissions

重印和许可

About this article

关于本文

Cite this article

引用本文

Shan, N., Shang, Y., He, Y.

。

et al.