文献速递重症综合管理与镇静镇痛消化、代谢与肾脏急症

机器学习和生物信息学识别脓毒症相关肾损伤中的凝血生物标志物

MACHINE LEARNING AND BIOINFORMATICS TO IDENTIFY COAGULATION BIOMARKERS IN SEPSIS-RELATED KIDNEY INJURY.
PubMed 文献 · Shock · 2025-01-01
一句话结论小鼠模型和临床队列发现C3、F3、Fgg、Serping1在脓毒症相关急性肾损伤中上调,多标志物组合诊断AUC为0.853。

该研究利用小鼠转录组数据集(GSE120879、GSE227623)整合生物信息学和机器学习,识别脓毒症相关急性肾损伤中的凝血相关差异表达基因,并在外部数据集(GSE142615)、qRT-PCR和免疫组化(盲肠结扎穿孔模型)中验证。四个核心基因C3、F3、Fgg、Serping1持续上调,qRT-PCR倍数变化7.4至23.6倍(P<0.05),F3蛋白表达经免疫组化确认(P<0.01)。免疫浸润分析显示T细胞/NK细胞浸润及PD-L1(CD274)与所有核心基因共表达(r=0.62-0.78,P<0.05)。临床队列(n=15)中,多标志物组合(纤维蛋白原、TAT、C3)诊断SA-AKI的AUC为0.853(95%CI:0.72-0.98)。

要点

  • 在盲肠结扎穿孔小鼠模型中,C3、F3、Fgg、Serping1四个凝血相关基因在SA-AKI中持续上调,qRT-PCR倍数变化7.4至23.6倍(P<0.05)。
  • F3蛋白表达经免疫组化确认(P<0.01)。
  • 免疫分析显示T细胞/NK细胞浸润,PD-L1(CD274)与所有核心基因共表达(r=0.62-0.78,P<0.05)。
  • 临床队列(n=15)中,多标志物组合(纤维蛋白原、TAT、C3)诊断SA-AKI的AUC为0.853(95%CI:0.72-0.98)。
  • 作者认为C3、F3、Fgg、Serping1是潜在的凝血-免疫生物标志物,具有经验证的诊断效用。

适用人群与场景

  • 小鼠脓毒症相关急性肾损伤模型(盲肠结扎穿孔)及临床队列(n=15)用于诊断性能评估。

研究设计与关键结果

  • 研究类型:生物信息学与机器学习分析结合实验验证。样本量:小鼠转录组数据集GSE120879、GSE227623,外部验证集GSE142615,临床队列n=15。人群:脓毒症相关急性肾损伤小鼠模型及临床队列。干预与对照:未提及。主要终点结果:四个核心基因C3、F3、Fgg、Serping1在SA-AKI中持续上调,qRT-PCR倍数变化7.4至23.6倍(P<0.05);F3蛋白表达经免疫组化确认(P<0.01)。关键次要终点:免疫分析显示T细胞/NK细胞浸润,PD-L1(CD274)与所有核心基因共表达(r=0.62-0.78,P<0.05);临床队列中多标志物组合(纤维蛋白原、TAT、C3)诊断SA-AKI的AUC为0.853(95%CI:0.72-0.98)。不良事件:未报告。

证据类型、等级与局限

  • 证据类型:生物信息学分析结合动物实验验证及小样本临床队列诊断性能评估。证据等级:原文未报告。局限:原文未报告局限性。

对床旁实践的改变

  • 作者认为C3、F3、Fgg、Serping1是潜在的凝血-免疫生物标志物,具有经验证的诊断效用,为SA-AKI早期检测提供了转化框架。

提炼说明

  • 本条依据论文摘要整理;摘要未涉及「诊断标准与关键阈值」「处置流程与药物剂量」「红旗征与不能漏的情况」「指南或共识出处」,故未列出这些栏目;全文请点「阅读原文」。
英文原文(摘要)· 点击展开

Background: Sepsis-associated acute kidney injury (SA-AKI) is a life-threatening complication with mortality rates exceeding 50%, yet its molecular drivers remain poorly defined. Dysregulated coagulation is increasingly implicated in SA-AKI pathogenesis through microvascular thrombosis and immune crosstalk, but kidney-specific coagulation biomarkers remain uncharacterized. Methods: Using murine (Mus musculus) transcriptomic datasets (GSE120879, GSE227623) from the NCBI GEO database, we integrated bioinformatics and machine learning to identify coagulation-related genes differentially expressed (DE-CRGs) in SA-AKI. Hub genes were validated via external datasets (GSE142615), qRT-PCR, and immunohistochemistry in a cecal ligation and puncture (CLP) mouse model. Immune infiltration and checkpoint correlations were analyzed using ImmuCellAI. Diagnostic performance was assessed in a clinical cohort (n = 15) via ROC curve. Results: Four hub DE-CRGs-C3, F3, Fgg, and Serping1-were consistently upregulated in murine SA-AKI (qRT-PCR fold-changes: 7.4- to 23.6-fold, P < 0.05). F3 protein expression was confirmed by immunohistochemistry ( P < 0.01). Immune profiling revealed T cell/NK cell infiltration and PD-L1 (CD274) co-expression with all hub genes (r = 0.62-0.78, P < 0.05). Clinically, a multimarker panel (fibrinogen, TAT, C3) achieved an AUC of 0.853 (95% CI: 0.72-0.98) for SA-AKI diagnosis. Conclusion: This study identifies C3, F3, Fgg, and Serping1 as potential novel coagulation-immune biomarkers for SA-AKI, with validated diagnostic utility. These findings bridge the critical knowledge gap between coagulation dysregulation and immune-mediated tubular injury in SA-AKI pathogenesis and provide a translational framework for early detection.

来源:Shock,查看原文。版权归原作者所有。

参考来源

  1. Liu L, Zhang T, Li L, et al. MACHINE LEARNING AND BIOINFORMATICS TO IDENTIFY COAGULATION BIOMARKERS IN SEPSIS-RELATED KIDNEY INJURY. Shock. 2025. PMID: 40208026. doi:10.1097/SHK.0000000000002600

供执业医师抢救与诊疗参考,不替代临床判断;剂量请结合体重、肝肾功能和本院药品规格复核,以指南原文为准。