目的 筛选直肠癌新辅助放化疗(CRT)疗效预测长链非编码RNA(lncRNA)分子标志物，分析参与CRT疗效调控相关信号通路，建立CRT疗效预测模型。方法 利用lncRNA芯片进行lncRNA差异表达检测，使用R软件Limma包在CRT反应组和CRT无反应组间对比筛选差异lncRNA(P＜0。05和|Log2FC|＞1)，进行分子标志物筛选。采用基因本体(GO)分析对差异表达基因进行功能分析，采用京都基因和基因组数据库(Kyoto Encyclopedia of Genes and Genomes，KEGG)对筛选的差异基因进行信号通路富集分析。进一步采用实时定量反转录聚合酶链式反应(qRT-PCR)检测98例样本。采用logistic回归构建CRT治疗预测模型。绘制受试者操作特征曲线(ROC)计算曲线下面积(AUC)以评价模型的判别区分能力。结果 CRT反应组中，823个lncRNA表达上调，216个lncRNA表达下调，449个基因表达上调，81个基因表达下调。新辅助放化疗相关上调排名前10的差异表达lncRNA 分别为 LUCAT1、LINC02356、HIF1A-AS2、Lnc-ZNF644-1、Lnc-ADAMTS12-3、LINC02356、Lnc-CLIC4-1、Lnc-PTX3-4、DARS-AS1、MIR210HG。下调排名前 10 的分别为 Lnc-COL6A3-2、Lnc-FBN1-2、Lnc-FOXA1-3、Lnc-KRTAP9-7-1、LINC00562、Lnc-NCS1-1、LINC00456、Lnc-FBLL1-2、USP2-AS1、Lnc-INPPL1-2。GO分析结果提示，差异基因主要富集在上皮细胞分化、中间丝、中间丝细胞骨架、突触后膜、颗粒分泌、细胞因子受体活性等分子生物学功能方面。KEGG富集分析提示，差异表达基因主要富集在HIF-1信号通路、Th17细胞分化、戊糖磷酸途径、精氨酸和脯氨酸代谢、果糖和甘露糖代谢、磷脂酶D信号通路、溶酶体等信号通路方面。logistic冋归模型显示，由LUCAT1、LINC02356、LINC00562三个lncRNA分子构成的预测模型具有较好的预测能力，AUC为0。887(95％CI 0。820～0。954)。模型回归方程 logit(p)=1。582×LINC00562-1。969xLINC02356-0。798x LUCAT1+4。357。模型的灵敏度为81。3％，特异度为84。0％。结论 直肠癌CRT反应良好和CRT无反应者间lncRNA分子存在明显的差异表达，由LUCAT1、LINC02356、LINC00562三个lncRNA分子构成的预测模型对CRT疗效具有较好的预测能力。
Construction of an lncRNA molecular prediction model for efficacy of neoadjuvant radiotherapy and chemotherapy for rectal cancer
Objective To screen long chain noncoding RNA(lncRNA)molecular markers for predicting the efficacy of neoadjuvant radiotherapy and chemotherapy(CRT)for rectal cancer,to analyze the signal pathways involved in the regulation of CRT efficacy,and to establish a model for predicting the efficacy of CRT.Methods The lncRNA microarray was used to analyze the differential expression of lncRNA.The R software Limma package was used to screen the differential lncRNA(P＜0.05 and log2fc＞1)between the CRT response group and the CRT non-response group,and the molecular markers were screened.GO analysis was used to analyze the function of differentially expressed genes,and Kyoto Encyclopedia of Genes and Genomes(KEGG)was used to analyze the signal pathway enrichment of the screened differentially expressed genes.Further,the real-time quantitative reverse transcription polymerase chain reaction(qRT-PCR)was used to detect its expression in the 98 collected samples.The logistic regression was used to construct a CRT prediction model.The area under the curve(AUC)was calculated by the receiver operating characteristic curve(ROC)to evaluate the discrimination ability of the model.Results In the CRT response group,823 lncRNA were up-regulated;216 lncRNA were down-regulated;449 genes were up-regulated;81 genes were down-regulated.The top 10 up-regulated lncRNA related to CRT were LUCAT1,LINC02356,HIF1A-AS2,Lnc-ZNF644-1,Lnc-ADAMTS12-3,LINC02356,Lnc-CLIC4-1,Lnc-PTX3-4,DARS-AS1,and M1R210HG,respectively.The top 10 down-regulated lncRNA were Lnc-COL6A3-2,Lnc-FBN1-2,Lnc-FOXA1-3,Lnc-KRTAP9-7-1,LINC00562,Lnc-NCS1-1,L1NC00456,Lnc-FBLL1-2,USP2-AS1,and Lnc-INPPL1-2,respectively.The GO analysis showed that the differential genes were mainly concentrated in the molecular biological functions,such as epithelial cell differentiation,intermediate filament,intermediate filament cytoskeleton,postsynaptic membrane,granule secretion,cytokine receptor activity,etc.The KEGG enrichment analysis showed that the differentially expressed genes were mainly enriched in the HIF-1 signal pathway,Th17 cell differentiation,pentose phosphate pathway,arginine and proline metabolism,fructose and mannose metabolism,phospholipase D signal pathway,lysosome,and other signal pathways.The logistic regression model showed that the model constructed by LUCAT1,LINC02356,and LINC00562 had good prediction ability,with an AUC of 0.887(95％CI 0.82-0.954).The model's regression equation was as below:logit(P)=1.582xLINC00562-1.969xLINC02356-0.798xLUCAT1+4.357.The sensitivity and specificity of the model were 81.3％and 84.0％,respectively.Conclusions There is a significant difference in the expression of lncRNA molecules between patients with good and poor CRT response.The prediction model composed of LUCAT1,LINC02356,and LINC00562 has good predictive ability for the efficacy of CRT.
Rectal cancerNeoadjuvant radiotherapy and chemotherapyLong chain noncoding RNAPrediction model
万方数据
维普
