基于TCGA和GEO分析m 6A调控基因在结直肠癌组织中的表达

Analysis of expression of m6A regulators in colorectal tumor samples based on TCGA and GEO databases

目的:探讨N6-甲基腺嘌呤(m^(6)A)在结直肠癌组织中表达及临床意义。 方法:2021年9月至2021年12月,利用癌症基因组图谱(TCGA)和基因表达综合数据库(GEO)对结直肠癌中m^(6)A相关调控因子的表达进行汇总分析,对不同临床病理特征及分子分型下m^(6)A调控因子的基因表达进行对比统计,对不同表达m^(6)A的结直肠癌病例做生存分析。 结果:在TCGA数据库中,胰岛素样生长因子2 mRNA绑定蛋白1(IGF2BP1),胰岛素样生长因子2 mRNA绑定蛋白3(IGF2BP3)和YTH结构域m6A RNA结合蛋白1(YTHDF1)在结直肠癌组织中较正常组织高表达(TCGA-COAD比IGF2BP3:12.80比204.46,logFC=4.00, P=0.003;YTHDF1:2 347.56比3 712.77,logFC=0.66, P < 0.001;TCGA-READ比IGF2BP1:6.20比359.32,logFC=5.82, P=0.007;YTHDF1:2 470.10比4 369.09,logFC=0.82, P=0.020);将TCGA和GEO数据库做交集后,发现甲基转移酶样14(METTL14)、YTH结构域m6A RNA结合蛋白3(YTHDF3)和α-酮戊二酸依赖的加双氧酶AlkB同源蛋白5(ALKBH5)均在结直肠癌组织中下调(TCGA-COAD比METTL14:1 051.56比711.40,logFC=-0.56, P < 0.001;YTHDF3:4 613.85比3 155.05,logFC=-0.55, P < 0.001;ALKBH5:4 250.10比2 555.55,logFC=-0.73, P < 0.001;TCGA-READ比METTL14:1 113.3比674.36,logFC=-0.72, P < 0.001;YTHDF3:5 034.30比3 331.95,logFC=-0.60, P=0.004;ALKBH5:4 902.20比2 529.71,logFC=-0.95, P < 0.001;GEO-CRC比METTL14:6.58比6.33,logFC=-0.06, P < 0.001;YTHDF3:6.28比6.20,logFC=-0.02, P=0.002;ALKBH5:5.07比4.98,logFC=-0.02, P < 0.001)。在不同分子分型结直肠癌中IGF2BP2、HNRNPC、TP53和YTHDF2在KRAS突变中低表达(IGF2BP2:48.53比44.04, t=2.64, P=0.008;HNRNPC:121.30比112.60, t=2.32, P=0.020;TP53:65.30比60.26, t=2.11, P=0.034;YTHDF2:54.07比51.19;t=1.97, P=0.048)。不同临床病理特征分析显示,IGF2BP1在淋巴结阳性(8.97比6.11,W=20 008, P=0.002)、存在远处转移(8.94比6.41,W=19 104, P=0.009)及Ⅲ~Ⅳ期(7.46比7.13,W=8 779, P=0.025)的结直肠癌中较对照组升高。ALKBH5高表达、高龄、存在脉管侵犯和病理分期晚与较短生存期显著相关(ALKBH5高表达比ALKBH5低表达:5.85年比NA, HR:1.63,95% CI:1.071~2.491, P=0.021。> 68岁比 < 68岁:6.78年比NA,HR:1.59,95% CI:1.049~2.422, P=0.047。Ⅲ~Ⅳ期比Ⅰ~Ⅱ期:4.55年比8.33年, HR:2.89,95% CI:1.895~4.425, P < 0.001。有静脉侵犯比无静脉侵犯:5.48年比NA, HR:2.82,95% CI:1.672~4.783, P < 0.001)与生存时间短具有明显相关性。 结论:m^(6)A中的部分调控因子与结直肠癌的生物学特征及预后相关。

Objective To detect the expression of N6-methyladenosine(m^(6)A)regulators in colorectal tumor samples and its clinical significance.Methods From September to December 2021,the data regarding the expression of m^(6)A regulators in colorectal tumor samples were collected using The Cancer Genome Atlas(TCGA)and Gene Expression Omnibus(GEO)and the expression of m^(6)A regulators was compared between different clinical pathological characteristics and between different molecular subtypes.Survival analysis was conducted in patients with colorectal cancer with different m^(6)A expression levels.Results In TCGA,insulin-like growth factor 2 mRNA binding protein 1(IGF2BP1),insulin-like growth factor 2 mRNA binding protein 3(IGF2BP3),and YTH N6-methyladenosine RNA binding protein 1(YTHDF1)were higher in colorectal tumor samples than normal tissue samples(TCGA-COAD:IGF2BP3:12.80 vs.204.46,logFC=4.00,P=0.003;YTHDF1:2347.56 vs.3712.77,logFC=0.66,P<0.001;TCGA-READ:IGF2BP1:6.20 vs.359.32,logFC=5.82,P=0.007;YTHDF1:2470.10 vs.4369.09,logFC=0.82,P=0.020).The intersection of TCGA and GEO databases showed that methyltransferase like 14(METTL14),YTH domain m^(6)A RNA binding protein 3(YTHDF3),andα-ketoglutarate dependent dioxygenase AlkB homolog 5(ALKBH5)were downregulated in colorectal tumor samples compared with normal tissue samples(TCGA-COAD:METTL14:1051.56 vs.711.40,logFC=-0.56,P<0.001;YTHDF3:4613.85 vs.3155.05,logFC=-0.55,P<0.001,ALKBH5:4250.10 vs.2555.55,logFC=-0.73,P<0.001;TCGA-READ vs.METTL14:1113.3 vs.674.36,logFC=-0.72,P<0.001;YTHDF3:5034.30 vs.3331.95,logFC=-0.60,P=0.004;ALKBH5:4902.20 vs.2529.71,logFC=-0.95,P<0.001;GEO-CRC vs.METTL14:6.58 vs.6.33,logFC=-0.06,P<0.001;YTHDF3:6.28 vs.6.20,logFC=-0.02,P=0.002;ALKBH5:5.07 vs.4.98,logFC=-0.02,P<0.001).In colorectal tumor samples of different molecular subtypes,IGF2BP2,HNRNPC,TP53 and YTHDF2 expression was low in KRAS(IGF2BP2:48.53 vs.44.04,t=2.64,P=0.008;HNRNPC:121.30 vs.112.60,t=2.32,P=0.020;TP53:65.30 vs.60.26,t=2.11,P=0.034;YTHDF2:54.07 vs.51.19;t=1.97,P=0.048).Different clinical pathological characteristics showed that IGF2BP1 was higher in colorectal cancer with positive lymph nodes(8.97 vs.6.11,W=20008,P=0.002),distant metastasis(8.94 vs.6.41,W=19104,P=0.009),or stageⅢ-Ⅳ(7.46 vs.7.13,W=8779,P=0.025)than normal tissue.ALKBH5 overexpression,advanced age,presence of vascular invasion,and late pathological staging were significantly related to a short survival period(high ALKBH5 expression vs.low ALKBH5 expression:5.85 years vs.not available,HR:1.63,95%CI:1.071-2.491,P=0.021;>68 years vs.<68 years:6.78 years vs.not available,HR:1.59,95%CI:1.049-2.422,P=0.047;stageⅢ-Ⅳvs.stageⅠ-Ⅱ:4.55 years vs.8.33 years,HR:2.89,95%CI:1.895-4.425,P<0.001;presence of vein invasion vs.absence of vein invasion:5.48 years vs.not available,HR:2.82,95%CI:1.672-4.783,P<0.001).Conclusion Some of the m^(6)A regulators are associated with the biological characteristics and prognosis of colorectal cancer.