基于加权共表达网络分析和 CIBERSORT 算法鉴定子宫平滑肌肉瘤相关枢纽基因和免疫细胞浸润模式。

BACKGROUND: While large-scale genomic analyses symbolize a precious attempt to decipher the molecular foundation of uterine leiomyosarcoma (ULMS), bioinformatics results associated with the occurrence of ULMS based totally on WGCNA and CIBERSORT have not yet been reported. This study aimed to screen the hub genes and the immune cell infiltration pattern in ULMS by bioinformatics methods. METHODS: Firstly, the GSE67463 dataset, including 25 ULMS tissues and 29 normal myometrium (NL) tissues, was downloaded from the public database. The differentially expressed genes (DEGs) were screened by the 'limma' package and hub modules were identified by weighted gene co-expression network analysis (WGCNA). Subsequently, gene function annotations were performed to investigate the biological role of the genes from the intersection of two groups (hub module and DEGs). The above genes were calculated in the protein-protein interaction (PPI) network to select the hub genes further. The hub genes were validated using external data (GSE764 and GSE68295). In addition, the differential immune cell infiltration between UL and ULMS tissues was investigated using the CIBERSORT algorithm. Finally, we used western blot to preliminarily detect the hub genes in cell lines. RESULTS: WGCNA analysis revealed a green-yellow module possessed the highest correlation with ULMS, including 1063 genes. A total of 172 DEGs were selected by thresholds set in the 'limma' package. The above two groups of genes were intersected to obtain 72 genes for functional annotation analysis. Interestingly, it indicated that 72 genes were mainly involved in immune processes and the Neddylation pathway. We found a higher infiltration of five types of cells (memory B cells, M0-type macrophages, mast cells activated, M1-type macrophages, and T cells follicular helper) in ULMS tissues than NL tissues, while the infiltration of two types of cells (NK cells activated and mast cells resting) was lower than in NL tissues. In addition, a total of five genes (KDR, CCL21, SELP, DPT, and DCN) were identified as the hub genes. Internal and external validation demonstrated that the five genes were over-expressed in NL tissues compared with USML tissues. Finally, the correlation analysis results indicate that NK cells activated and mast cells activated positively correlated with the hub genes. However, M1-type macrophages had a negative correlation with the hub genes. Moreover, only the DCN may be associated with the Neddylation pathway. CONCLUSION: A series of evidence confirm that the five hub genes and the infiltration of seven types of immune cells are related to USML occurrence. These hub genes may affect the occurrence of USML through immune-related and Neddylation pathways, providing molecular evidence for the treatment of USML in the future.

背景：虽然大规模基因组分析标志着破译子宫平滑肌肉瘤（ULMS）分子基础的宝贵尝试，但基于 WGCNA 和 CIBERSORT 的 ULMS 发生的生物信息学结果尚未报道。本研究旨在通过生物信息学方法筛选 ULMS 的枢纽基因和免疫细胞浸润模式。

方法：首先，从公共数据库中下载了包含 25 个 ULMS 组织和 29 个正常子宫肌层（NL）组织的 GSE67463 数据集。使用'limma'包筛选差异表达基因（DEGs），并通过加权基因共表达网络分析（WGCNA）鉴定枢纽模块。随后，对两组（枢纽模块和 DEGs）的基因进行基因功能注释，以研究基因的生物学作用。在蛋白质-蛋白质相互作用（PPI）网络中计算上述基因，以进一步选择枢纽基因。使用外部数据（GSE764 和 GSE68295）验证枢纽基因。此外，使用 CIBERSORT 算法研究 UL 和 ULMS 组织之间的差异免疫细胞浸润。最后，我们使用 Western blot 初步检测细胞系中的枢纽基因。

结果：WGCNA 分析显示，与 ULMS 相关性最高的是一个绿色-黄色模块，包含 1063 个基因。通过'limma'包设置的阈值共筛选出 172 个 DEGs。上述两组基因相交，获得 72 个用于功能注释分析的基因。有趣的是，结果表明 72 个基因主要参与免疫过程和 Neddylation 途径。我们发现 ULMS 组织中五种细胞（记忆 B 细胞、M0 型巨噬细胞、激活的肥大细胞、M1 型巨噬细胞和滤泡辅助性 T 细胞）的浸润高于 NL 组织，而两种细胞（NK 细胞激活和静止的肥大细胞）的浸润低于 NL 组织。此外，共鉴定出 5 个基因（KDR、CCL21、SELP、DPT 和 DCN）为枢纽基因。内部和外部验证表明，与 USML 组织相比，这 5 个基因在 NL 组织中表达上调。最后，相关性分析结果表明，NK 细胞激活和激活的肥大细胞与枢纽基因呈正相关。然而，M1 型巨噬细胞与枢纽基因呈负相关。此外，只有 DCN 可能与 Neddylation 途径有关。

结论：一系列证据证实，这 5 个枢纽基因和 7 种免疫细胞浸润与 USML 的发生有关。这些枢纽基因可能通过免疫相关和 Neddylation 途径影响 USML 的发生，为未来 USML 的治疗提供分子证据。