检测肾移植受者术后血浆外泌体miR-21、miR-210和miR-4639表达变化，分析外泌体miR-21、miR-210和miR-4639单独及联合对肾移植术后并发慢性移植肾肾病(CAN)的诊断价值。

回顾性分析2018年1月至2019年1月苏州大学附属第三医院泌尿外科实施的同种异体肾移植受者临床资料，最终纳入34例受者，根据肾移植术后是否发生CAN将其分为CAN组及对照组。采用凝胶排阻色谱法提取血浆外泌体，采用Nanosight NS300分析外泌体粒径，采用蛋白质印迹法(WB)分析外泌体表面标志物(CD63和Alix)表达情况。采用卡方检验比较CAN组和对照组受者性别比例。采用成组t检验比较两组受者移植前年龄、末次血清肌酐、血清尿素氮和估算肾小球滤过率(eGFR)。采用受试者工作特征(ROC)曲线评价血浆外泌体miR-210、miR-21和miR-4639对肾移植术后并发CAN的诊断效能。P<0.05为差异有统计学意义。

CAN组(n＝18例)和对照组(n＝16例)受者性别以及移植前年龄、末次血清肌酐、血清尿素氮和eGFR差异均无统计学意义(χ²＝0.04、t＝0.86、-1.84、-1.83和0.85，P均>0.05)。透射电镜、Nanosight NS300及WB检测结果均提示提取样本为血浆外泌体。CAN组与对照组血浆外泌体miR-210、miR-21和miR-4639相对表达量差异均有统计学意义(t＝4.13、3.38和2.33，P均<0.05)。miR-210预测肾移植术后并发CAN的ROC曲线下面积为0.854(95%CI：0.730～0.979，P<0.05)，当截断值＝1.320时，敏感度为66.7%，特异度为93.8%。miR-21预测肾移植术后并发CAN的ROC曲线下面积为0.774(95%CI：0.618～0.931，P<0.05)，当截断值＝1.243时，敏感度为55.6%，特异度为93.8%。miR-4639预测肾移植术后并发CAN的ROC曲线下面积为0.670(95%CI：0.482～0.859，P<0.05)，当截断值＝0.936，敏感度为66.7%，特异度为75.0%。随后，构建基于miR-210、miR-21和miR-4639 3个指标的联合诊断模型，回归方程z＝5.293×[miR-210]＋5.046×[miR-21]＋0.433×[miR-4639]-13.373，联合预测概率值p＝e^z/(1＋e^z)。miR-210、miR-21和miR-4639联合预测肾移植术后并发CAN的ROC曲线下面积为0.938(95%CI：0.860～1.015，P<0.05)，当截断值＝0.587，敏感度为83.33%，特异度为93.75%。当联合预测值为0.587时，CAN组有83.3%(15/18)的个体被联合预测模型诊断出阳性结果，而对照组有93.8%(15/16)的个体被联合预测模型诊断出阴性结果，表明该联合预测模型有较好的诊断价值。

miR-210、miR-21和miR-4639组成的miRNA阵列可能可以用于早期诊断肾移植术后并发CAN。

To detect the changes in the expression of plasma exosomal miR-21, miR-210 and miR-4639 in renal transplant recipients, and to analyze the diagnostic value of miR-21, miR-210 and miR-4639, and combined use of them for chronic allograft nephropathy (CAN).

The clinical data of kidney transplant recipients in the Department of Urology, the Third Affiliated Hospital of Soochow University from January 2018 to January 2019 were retrospectively analyzed. Thirty-four kidney transplant recipients were enrolled and divided into CAN group and control group. Plasma exosomes were isolated by gel exclusion chromatography, and the particle size analysis of exosomes was detected by Nanoparticle NS300. Expression of exosome surface markers (CD63 and Alix) were analyzed by western blotting (WB). The chi-square test was used to compare the gender ratio of the CAN group and the control group. Age, serum creatinine, urea nitrogen and estimated glomerular filtration rate (eGFR) (last time before transplantation) of the 2 groups of recipients were compared by group t test. The receiver operating characteristic (ROC) curve was used to evaluate the diagnostic efficacy of plasma exosomal miR-210, miR-21 and miR-4639 for CAN after renal transplantation. P<0.05 was considered statistically significant.

There was no statistically significant differences in gender, age, serum creatinine, urea nitrogen and eGFR before transplantation between the CAN group (n=18) and the control group (n=16) (χ²=0.04, t=0.86, -1.84, -1.83 and 0.85, P all >0.05). Transmission electron microscopy, Nanosight NS300 and WB analysis indicated that the extracted samples were plasma exosomes. The relative expressions of miR-210, miR-21 and miR-4639 between the 2 groups had statistical significance (t=4.13, 3.38 and 2.33, P all<0.05). The area under the ROC curve (AUC) for the prediction of CAN by miR-210 was 0.854(95%CI: 0.730-0.979, P<0.05). When the cut-off value was 1.320, the sensitivity was 66.7%, and the specificity was 93.8%. The AUC for the prediction of CAN by miR-21 was 0.774(95%CI: 0.618-0.931, P<0.05). When the cut-off value was 1.243, the sensitivity was 55.6%, and the specificity was 93.8%. The AUC for the prediction of CAN by miR-4639 was 0.670(95%CI: 0.482-0.859, P<0.05). When the cut-off value was 0.936, the sensitivity was 66.7%, and the specificity was 75.0%. A joint diagnostic model based on 3 indicators was constructed, and the regression equation was z=5.293×[miR-210]+ 5.046×[miR-21]+ 0.433×[miR-4639] -13.373. The joint prediction value p=e^z/(1+ e^z). The AUC of combined miRNAs to predict the complications of CAN after renal transplantation was 0.938 (95%CI: 0.860-1.015, P<0.05). When the cut-off value was 0.587, the sensitivity was 83.3% and specificity was 93.8%. Moreover, 83.3% (15/18) of the individuals in the CAN group were diagnosed as positive by the joint prediction model, while 93.8% (15/16) of the individuals in the control group were diagnosed as negative by the joint prediction model when the cutoff value 0.587 was applied. The results indicate that the joint prediction model has good diagnostic value.

Joint diagnostic model of plasma exosomal miR-210, miR-21 and miR-4639 may become a biomarker for early noninvasive diagnosis of CAN after renal transplantation.