Department of Anesthesiology, The First Affiliated Hospital, Sun Yat-sen University, Guangzhou, China.
Zhongshan School of Medicine, Sun Yat-sen University, Guangzhou, China.
Front Immunol. 2023 Jun 2;14:1078055. doi: 10.3389/fimmu.2023.1078055. eCollection 2023.
There is still a lack of specific indicators to diagnose idiopathic pulmonary fibrosis (IPF). And the role of immune responses in IPF is elusive. In this study, we aimed to identify hub genes for diagnosing IPF and to explore the immune microenvironment in IPF.
We identified differentially expressed genes (DEGs) between IPF and control lung samples using the GEO database. Combining LASSO regression and SVM-RFE machine learning algorithms, we identified hub genes. Their differential expression were further validated in bleomycin-induced pulmonary fibrosis model mice and a meta-GEO cohort consisting of five merged GEO datasets. Then, we used the hub genes to construct a diagnostic model. All GEO datasets met the inclusion criteria, and verification methods, including ROC curve analysis, calibration curve (CC) analysis, decision curve analysis (DCA) and clinical impact curve (CIC) analysis, were performed to validate the reliability of the model. Through the Cell Type Identification by Estimating Relative Subsets of RNA Transcripts algorithm (CIBERSORT), we analyzed the correlations between infiltrating immune cells and hub genes and the changes in diverse infiltrating immune cells in IPF.
A total of 412 DEGs were identified between IPF and healthy control samples, of which 283 were upregulated and 129 were downregulated. Through machine learning, three hub genes () were screened. We confirmed their differential expression using pulmonary fibrosis model mice evaluated by qPCR, western blotting and immunofluorescence staining and analysis of the meta-GEO cohort. There was a strong correlation between the expression of the three hub genes and neutrophils. Then, we constructed a diagnostic model for diagnosing IPF. The areas under the curve were 1.000 and 0.962 for the training and validation cohorts, respectively. The analysis of other external validation cohorts, as well as the CC analysis, DCA, and CIC analysis, also demonstrated strong agreement. There was also a significant correlation between IPF and infiltrating immune cells. The frequencies of most infiltrating immune cells involved in activating adaptive immune responses were increased in IPF, and a majority of innate immune cells showed reduced frequencies.
Our study demonstrated that three hub genes (, ) were associated with neutrophils, and the model constructed with these genes showed good diagnostic value in IPF. There was a significant correlation between IPF and infiltrating immune cells, indicating the potential role of immune regulation in the pathological process of IPF.
特发性肺纤维化（IPF）的诊断仍缺乏特异性指标，其免疫反应的作用也难以捉摸。本研究旨在寻找用于诊断 IPF 的关键基因，并探讨 IPF 的免疫微环境。
我们利用 GEO 数据库，鉴定出 IPF 与对照肺组织样本之间的差异表达基因（DEGs）。通过 LASSO 回归和 SVM-RFE 机器学习算法，我们鉴定出关键基因。在博来霉素诱导的肺纤维化模型小鼠和由五个合并 GEO 数据集组成的 Meta-GEO 队列中，进一步验证这些基因的差异表达。然后，我们使用这些基因构建诊断模型。所有 GEO 数据集均符合纳入标准，采用 ROC 曲线分析、校准曲线（CC）分析、决策曲线分析（DCA）和临床影响曲线（CIC）分析等验证方法，验证模型的可靠性。通过基于 RNA 转录本相对亚群估计的细胞类型鉴定算法（CIBERSORT），我们分析了浸润性免疫细胞与关键基因之间的相关性，以及 IPF 中不同浸润性免疫细胞的变化。
在 IPF 与健康对照样本之间共鉴定出 412 个 DEGs，其中 283 个基因上调，129 个基因下调。通过机器学习，筛选出 3 个关键基因（、和）。我们通过 qPCR、western blot 和免疫荧光染色评估肺纤维化模型小鼠，以及 Meta-GEO 队列的分析，验证了这些基因的差异表达。这三个关键基因的表达与中性粒细胞有很强的相关性。然后，我们构建了一个用于诊断 IPF 的诊断模型。在训练队列和验证队列中，曲线下面积分别为 1.000 和 0.962。对其他外部验证队列的分析，以及 CC 分析、DCA 和 CIC 分析也表明了很好的一致性。IPF 与浸润性免疫细胞之间也存在显著相关性。参与激活适应性免疫反应的大多数浸润性免疫细胞的频率在 IPF 中增加，而大多数固有免疫细胞的频率降低。
本研究表明，三个关键基因（、和）与中性粒细胞相关，由这些基因构建的模型在 IPF 诊断中具有良好的诊断价值。IPF 与浸润性免疫细胞之间存在显著相关性，提示免疫调节在 IPF 的病理过程中可能发挥作用。
