目的：含蒽环类药物化疗方案所致肝损伤（anthracycline-induced liver injury, AILI）是乳腺癌术后辅助化疗的严重并发症之一。外泌体作为细胞间通信的载体，在药物性肝损伤（drug-induced liver injury, DILI）中发挥重要作用。本研究探讨了外泌体内容物作为AILI标志物的潜力。

方法：本研究共招募70名化疗后乳腺癌患者。在分离出血浆来源的外泌体后，使用RNA测序揭示外泌体mRNA、circRNA和miRNA表达谱，并对非肝损伤组和肝损伤组进行基因差异表达分析。在mRNA的研究中，使用DAVID数据库对差异表达的mRNA进行GO和KEGG富集分析。使用STRING数据库建立差异表达mRNA的蛋白-蛋白相互作用网络（PPI网络），并通过Cytoscape软件鉴定出在AILI中发挥重要作用的外泌体mRNA。在circRNA的研究中，qRT-PCR用于验证差异表达的circRNA。CircInteractome数据库和CSCD数据库用于预测circRNA的靶miRNA，miRTarBase数据库用于检索miRNA的靶mRNA。靶mRNA与CTD数据库中的DILI相关基因取交集后，根据相互作用关系，构建circRNA-miRNA-mRNA调控网络。使用DAVID数据库对circRNA间接作用的mRNA进行GO和KEGG富集分析，以预测circRNA发挥的调控功能。在miRNA的研究中，使用miRTarBase和miRDB数据库预测差异表达miRNA的靶基因。靶基因与DILI相关基因的交集被定义为AILI相关靶基因。使用DAVID数据库对AILI相关靶基因进行GO和KEGG富集分析。使用STRING数据库建立AILI相关靶基因的PPI网络，通过Cytoscape软件鉴定出在AILI中发挥重要作用的外泌体miRNA。

结果：非肝损伤组和肝损伤组间差异表达的mRNA有443个。富集分析结果显示，差异表达的mRNA显著富集于Notch信号通路、白细胞介素-10产生的正向调节、线粒体膜电位的调控、对活性氧的反应、铜离子结合、谷胱甘肽代谢、T细胞受体信号通路和细胞色素P450代谢异源物质等肝损伤相关通路。在PPI网络筛选出的10个hub基因中，ALB、KRT8和VEGFA是DILI相关基因。非肝损伤组和肝损伤组间差异表达的circRNA有34个。PCR结果显示，circ_0005130和circ_0094333在患者中的表达情况与测序结果一致。经生物信息学分析，共预测出13条circRNA-miRNA-mRNA调控网络，circRNA可能经这些途径参与调控细胞对有毒物质的反应、氧化还原过程等生物学功能。非肝损伤组和肝损伤组间差异表达的miRNA有30个，miRNA的AILI相关靶基因有79个。AILI相关靶基因在NOD样受体信号通路、HIF-1信号通路和FoxO信号通路中显著富集。在PPI网络中，IL-6和SOD2是最为重要的hub基因，可分别通过激活免疫反应和氧化应激反应参与AILI的发生发展。在miRNA-hub基因网络中，miR-1-3p可作用于多个hub基因，发挥重要调控作用。

结论：ALB、KRT8、VEGFA、circ_0005130、circ_0094333、miR-1-3p可能在AILI中发挥重要功能，这些分子有潜力成为AILI诊断或预后预测的生物标志物。AILI的发病机制可能为氧化应激、抗氧化酶水平降低、线粒体功能障碍、炎症反应等生物学过程的激活。

Objective: Anthracycline-induced liver injury (AILI) is one of the serious complications of adjuvant chemotherapy for postoperative breast cancer patients. Exosomes, as signaling molecules in intercellular communication, play essential roles in drug-induced liver injury (DILI). This study investigated the potential of exosomal contents as AILI biomarkers.

Methods: A total of 70 post-chemotherapy breast cancer patients were recruited in this study. After isolated plasma-derived exosomes, RNA sequencing revealed the exosomal mRNA, circRNA, and miRNA profiles and differentially expressed mRNA, circRNA, and miRNA were identified between the non-liver injury group and liver injury group. In the study of mRNA, GO and KEGG enrichment analyses of differentially expressed mRNAs were performed using the DAVID database. The protein-protein interaction network (PPI network) of differentially expressed mRNAs was established by the STRING database, and exosomal mRNAs that play an important role in AILI were identified by Cytoscape software. In the study of circRNA, qRT-PCR was applied to verify the differentially expressed circRNAs. The CircInteractome database and the CSCD database were employed to predict the target miRNA of circRNA, and the miRTarBase database was used to query the target mRNA of miRNA. The target mRNAs were intersected with DILI-related genes in the CTD database, and the circRNA-miRNA-mRNA regulatory networks were constructed according to the interaction relationship. GO and KEGG enrichment analyses were performed on mRNAs using the DAVID database to predict the regulatory function of circRNA. In miRNA research, the miRTarBase database and the miRDB database were used to predict the target genes of differentially expressed miRNAs. The intersection of target genes and DILI-related genes were defined as AILI-related target genes. GO and KEGG enrichment analyses of AILI-related target genes were performed using the DAVID database. The STRING database was used to establish the PPI network of AILI-related target genes, and the Cytoscape software was used to identify exosomal miRNAs that play an important role in AILI.

Results: There were 443 mRNAs differentially expressed between the non-liver injury group and the liver injury group. The results of enrichment analysis showed that differentially expressed mRNAs were significantly enriched in Notch signaling pathway, positive regulation of interleukin-10 production, regulation of mitochondrial membrane potential, response to reactive oxygen species, copper ion binding, glutathione metabolism, T cell receptor signaling pathway, metabolism of xenobiotics by cytochrome P450 and other liver injury-related pathways. Among the 10 hub genes screened from the PPI network, ALB, KRT8 and VEGFA were DILI-related genes. There were 34 circRNAs differentially expressed between the non-liver injury group and the liver injury group. PCR results showed that the expressions of circ_0005130 and circ_0094333 were consistent with the sequencing results. A total of 13 circRNA-miRNA-mRNA regulatory networks were predicted by bioinformatics analysis, through which circRNAs can participate in the regulation of response to toxic substance, oxidation-reduction process and other biological functions. There were 30 miRNAs differentially expressed between the non-liver injury group and the liver injury group, and there were 79 AILI-related target genes of miRNA. AILI-related target genes were significantly enriched in NOD-like receptor signaling pathway, HIF-1 signaling pathway and FoxO signaling pathway. In the PPI network, IL-6 and SOD2 are the most important hub genes, which can participate in the occurrence and development of AILI by activating immune response and oxidative stress response, respectively. In the miRNA-hub gene network, miR-1-3p can bind multiple hub genes and play an important regulatory role.

Conclusions: ALB, KRT8, VEGFA, circ_0005130, circ_0094333 and miR-1-3p may play an important role in AILI, and these molecules have the potential to be biomarkers for diagnosis or prognosis prediction of AILI. The pathogenesis of AILI may be the activation of biological processes such as oxidative stress, reduced levels of antioxidant enzymes, mitochondrial dysfunction, and inflammation.