目的 探讨免疫细胞分化对脓毒性ARDS肺泡-毛细血管屏障损伤的影响.方法 基于转录组数据构建脓毒性ARDS的WGCNA网络,并筛选ARDS相关基因(ARDS-related genes,ARGs).基于单细胞测序数据构建脓毒性ARDS分化轨迹和细胞通讯,并筛选免疫分化相关基因(immunodifferentiation-related genes,IDRGs).Lasso 回归分析构建 ARDS 的免疫相关风险评分(risk Score,RS).ESTIMATE、CIBERSORT 和 ssGSEA 评估免疫微环境.Metascape、GSVA 和 GO 富集分析展示信号通路和生物学过程.结果 免疫细胞、成纤维细胞和内皮细胞通过24条信号通路进行细胞通讯.由DSTN、SNRPA和FGL2组成的RS在正常、单纯脓毒症和脓毒性ARDS的患者中差异表达,涉及免疫细胞、内皮细胞和成纤维细胞的分化,影响脓毒性ARDS的免疫浸润和免疫功能.脓毒性ARDS相关免疫细胞包含记忆B细胞、浆细胞、CD8+T和M0.DSTN与M0负相关(r=-0.29,P＜0.05),SNRPA 与 CD8+T 正相关(r=0.28,P＜0.05),FGL2 与记忆 B 细胞正相关(r=0.32,P＜0.05).RS组间差异表达的免疫细胞包括CD4幼稚型T细胞、CD4记忆激活T细胞、调节T细胞、γ-δ T细胞、M0、激活的树突状细胞.富集分析提示DSTN、SNRPA和FGL2的差异表达影响了免疫细胞的分化、免疫功能的激活、抗原的呈递,以及肌动蛋白丝的解聚/切断.结论 DSTN、SNRPA和FGL2通过调控免疫细胞、成纤维细胞和内皮细胞的分化,影响脓毒性ARDS的免疫性肺泡-毛细血管屏障损伤,是预测脓毒性ARDS进展的潜在标志物.

目的 基于癌症基因组图谱(the cancer genome atlas,TCGA)数据库构建肝细胞癌(hepatocellular carcinoma,HCC)双硫死亡相关基因(disulfidptosis-related genes,DRGs)预后风险模型及评价.方法 通过生物信息学方法分析TCGA数据库中 371 例HCC样本及 50 例癌旁样本中 15 个DRGs的表达情况,并进行基因本体(gene ontology,GO)功能注释和京都基因和基因组百科全书(Kyoto encyclopedia of genes and genomes,KEGG)富集分析、Kaplan-Meier(KM)生存分析;通过单因素COX回归分析筛选出有统计学意义的DRGs,通过LASSO回归分析及多因素COX回归分析筛选出关键DRGs构建预后风险模型,并根据风险评分将HCC患者分为高风险组和低风险组,制作KM生存曲线和时间依赖的受试者工作特征(receiver operator characteristic,ROC)曲线进行验证评价.结果 与癌旁样本相比,HCC样本 15 个DRGs中FLNA,MYH9,TLN1,ACTB,MYL6,CAPZB,DSTN,ACTN4,SLC7A11,INF2,CD2AP,PDLIM1 和FLNB均表达上调,且差异具有统计学意义(t=1 793～6 310,均P<0.001);经GO功能注释和KEGG富集分析显示,DRGs主要与肌动蛋白细胞骨架和细胞黏附相关的生物过程或途径密切相关.经KM生存分析结果显示,SLC7A11,INF2,CD2AP,MYL6,ACTB高表达组生存率低于低表达组[HR=1.46(1.03～2.07)～1.93(1.36～2.75),均P<0.05].通过单因素COX回归分析、LASSO分析及多因素COX回归分析构建预后风险模型riskscore=(0.247×SLC7A11)+(0.289×INF2)+(0.076×CD2AP)+(0.06×MYL6);计算样本的风险评分,风险评分越高,预后不良的HCC患者人数越多;KM生存分析显示高风险组的总生存率比低风险组低;1,3,5 年的AUC值分别为 0.709,0.661 和 0.648;通过多因素COX回归分析表明SLC7A11[HR=1.832(1.274～2.636),P=0.001]是独立的预后危险因素.结论 四个DRGs构建的预后风险模型在预测HCC患者预后情况有一定的作用.

Objective To explore the candidate genes that play significant roles in the interconnection between abdominal aortic aneurysm (AAA) and type 2 diabetes mellitus (DM). Methods We used the Biomedical Discovery Support System (BITOLA) to screen out the candidate intermediate molecular (CIM) "Gene or Gene Product" that are related to AAA and DM. The dataset of GSE13760, GSE7084, GSE57691, GSE47472 were used to analyze the differentially expressed genes (DEGs) of AAA and DM compared to the healthy status. We used the online tool of Venny 2.1 assisted by manual checking to identify the overlapped DEGs with the CIMs. The Human eFP Browser was applied to examine the tissue specific expression levels of the detected genes in order to recognize strong expressed genes in both human artery and pancreatic tissue. Results There were 86 CIMs suggested by the closed BITOLA system. Among all the DEGs of AAA and DM, 8 genes in GSE7084 (ISG20, ITGAX, DSTN, CCL5, CCR5, AGTR1, CD19, CD44) and 2 genes in GSE13760 (PSMD12, FAS) were found to be overlapped with the 86 CIMs. By manual checking and comparing with tissue- specific gene data through Human eFP Browser, the gene PSMD12 (proteasome 26S subunit, non-ATPase 12) was recognized to be strongly expressed in both the aorta and pancreatic tissue. Conclusion We proposed a hypothesis through text mining that PSMD12 might be involved or potentially involved in the interconnection between AAA and DM, which may provide a new clue for studies on novel therapeutic strategies for the two diseases.