【作者】 徐俊；

【导师】 陈卫昌；

【作者基本信息】 苏州大学 ， 消化内科（专业学位）， 2021， 博士

【摘要】 第一部分基于肠道菌群的结直肠癌失衡指数分析目的:结直肠癌(Colorectal cancer,CRC)患者存在肠道菌群失衡,其发生是一个涉及多个细菌种类或类群组成改变的复杂过程。如果能将这种复杂的变化提炼为一个简单的指标,将有助于推动肠道菌群相关研究结果的临床应用。本研究旨在通过16S rRNA测序建立失衡指数模型,以评估与CRC相关的肠道菌群失衡情况。方法:收集自2017年1月至2018年7月在苏州大学附属第一医院行肠镜检查的129例志愿者粪便样本,其中CRC组70例,对照组59例,所有样本均于手术或内镜操作前采集。通过16S rRNA高通量测序分析样本中的微生物组成及总体多样性,比较CRC组和对照组肠道菌群组成差异,并进一步分析不同病变部位、不同病理类型CRC患者的菌群特征。基于CRC组相对丰度升高或降低的菌群建立加权失衡指数(Weighted dysbiosis index of cancer,WDIc)模型,以此检测另一单独队列的 123 例样本,同时进一步验证了来自不同作者发表的3个公共数据集中的396例样本。结果:共鉴定出7个丰度差异显著的类群:与对照组相比,具核梭杆菌、微小微单胞菌、消化链球菌、Clostridiales Family Ⅺ以及卟啉单胞菌在CRC组明显增多,而假丁酸弧菌和唾液链球菌在CRC组明显减少。同时发现,不同病变部位、不同病理类型CRC患者存在特征性的肠道菌群组成差异。以CRC组7个差异类群建立的WDIc评分显著高于对照组,且与肿瘤大小和TNM分期正相关。WDIc可以有效区分CRC组和对照组,其受试者工作特征曲线(Receiver operating characteristic curve,ROC)下面积(Area under ROC curve,AUC)为 0.83,敏感性为 71.4%,特异性为83.1%。以此模型检测第二队列患者时的敏感性为71.0%,特异性为87.0%,而在三个公共数据集中进一步验证时同样发现其可以有效区分CRC组和对照组。结论:我们建立了一个可以衡量CRC患者肠道菌群失衡的指数,其具有较好的区分CRC患者和健康人群的能力,可能有助于促成肠道菌群检测在CRC诊治中的潜在临床应用。第二部分肠道菌群变化与结直肠腺瘤的相关性研究目的:结直肠腺瘤(Colorectal adenoma,CRA)是经典的CRC癌前病变,如能早期诊断和治疗,将极大改善患者预后。但临床常用的粪便隐血试验(Fecal occult blood test,FOBT)、血清 CEA(Carcinoembryonic antigen)等筛查指标对 CRA 判别能力较低,尚缺乏无创且高效的早期检测手段。本研究旨在基于16S rRNA测序基础上,寻找一个新的指标用于结直肠早期病变的筛查。方法:收集自2017年1月至2018年7月在苏州大学附属第一医院行肠镜检查的155例志愿者粪便样本,其中CRA组96例(包括进展期腺瘤52例,非进展期腺瘤44例),对照组59例。均接受FOBT和血清CEA检测。通过16S rRNA高通量测序比较CRA和对照组的肠道菌群组成,并分析CRA组内非进展期和进展期不同阶段的菌群构成差异,同时进一步探讨在结直肠肿瘤发展过程中肠道宏基因组的功能变化。基于CRA与对照组差异菌株构建加权失衡指数(Weighted dysbiosis index of adenoma,WDIa)模型,对比分析此模型与FOBT、血清CEA在CRA临床检测中的效能。结果:CRA患者具有更高的体重指数(Body mass index,BMI)(P=0.014),腺瘤大小与进展期风险呈显著正相关(P=9.1×10-18)。相比对照组,CRA组肺炎克雷伯杆菌、粪副拟杆菌、梭杆菌属显著增多,而毛螺菌科、布劳特氏菌属、放线菌门、双歧杆菌属、罗氏杆菌属、链球菌属、假丁酸弧菌属、溃疡梭杆菌显著减少。普雷沃菌属和拟杆菌属在非进展期腺瘤组中显著增多,而厌氧消化链球菌、Lachnoclostridium在进展期腺瘤组中明显增多。同时发现以氨基酸代谢,心血管、神经、内分泌系统,过氧化物酶体增殖物激活受体(Peroxisome proliferators-activatedreceptors,PPARs)和P53等为代表的30条京都基因与基因组百科全书(Kyoto encyclopedia of genes and genomes,KEGG)信号通路在结直肠肿瘤发展的不同阶段存在明显的差异表达。比较WDIa、FOBT和血清CEA的检验效能,发现FOBT对于CRA的敏感性为45.5%,特异性为91.5%,误诊率为8.5%,漏诊率为54.5%。CEA对CRA的诊断敏感性为52.3%,特异性70.6%,AUC为0.614,而WDIa的敏感性为88.1%,特异性为54.2%,AUC 为 0.756。结论:在结直肠肿瘤发展的CRA阶段就存在着特征性的肠道菌群构成和宏基因组功能变化。根据CRA和对照组核心差异菌株构建模型的灵敏度明显优于FOBT及血清CEA检测,能更快速便捷的早期识别和诊断CRA,或许有望成为结直肠肿瘤早期筛查的潜在标志物。第三部分微小微单胞菌在结直肠腺瘤-癌序列中的丰度变化目的:目前认为微小微单胞菌(Parvimonas micra,P.micra)与CRC密切相关。然而,它与CRA的关系及其在结直肠肿瘤发生中的作用尚不明确。本研究旨在通过分析一个苏州地区队列和4个公共测序数据集中P.micra在腺瘤到癌的发展进程中的丰度变化,探讨P.micra与CRA和CRC的关系。方法:运用定量 PCR(Quantitative polymerase chain reaction,qPCR)方法对 277例自2017年1月至2019年3月间在苏州大学附属第一医院行肠镜检查的门诊患者(包括CRA 128例,CRC 66例,健康对照83例)粪便样本进行检测,探讨在腺瘤到癌进程中肠道内P.micra的丰度变化情况。同时,在4个公共的16S rRNA测序数据集中,分析了 596名受试者(包括CRA 253例,CRC 158例,健康对照185例)粪便样本的P.micra相对丰度。结果:CRA组P.micra丰度与健康对照组相似(P=0.2),而CRC组P.micra丰度显著增加(P=8.2×10-11)。P.micra丰度水平可以有效区分CRC和健康对照者,但对于CRA的区分能力相对有限,其对CRC患者ROC曲线下面积为0.867,CRA患者为0.554。在所有的4个公共测序数据集中均发现了上述相同的P.micra丰度变化模式,即P.micra在健康对照组和CRA患者中丰度相对较低,而在CRC患者中则显著升高。结论:P.micra与CRC而非CRA密切相关,并有可能成为CRC的诊断标志物。此外,P.micra作为一种机会致病菌,可能促进了 CRC的发展,但其在CRA阶段的作用可能相对有限。更多还原

【Abstract】 Part Ⅰ:A Weighted Dysbiosis Index for Colorectal Cancer Analysis Based on Gut MicrobiomeObjective:Occurrence of gut microbial dysbiosis in colorectal cancer(CRC)is a complex process involving compositional alterations of multiple bacterial species or taxa.Condensing this complex changes into a simple index would facilitate clinical application of gut microbiome data.The study aim was to build a single numeric index to measure gut dysbiosis associated with CRC based on 16S rRNA sequencing data.Methods:Stool samples were collected froml29 patients belonging to CRC group(70 cases)and healthy group(59 cases)prior to colon resection surgery or endoscopic examination.Microbial composition and diversity were analyzed by 16S rRNA sequencing to compare the composition of gut microbiota between the control and CRC group,as well as among the patients with different pathological lesions and types of CRC.Based on the relative abundance of CRC enriched or depleted taxa,a weighted dysbiosis index of cancer(WDIc)model was built,tested with 123 samples from a separate cohort of patients,and further validated with 396 samples from three datasets published by various authors.Results:A signature of 7 taxa was identified:Fusobacterium nucleatum,Parvimonas,Peptostreptococcus,Clostridiales Family Ⅺ and Porphyromonas were significantly enriched in CRC group while Pseudobutyrivibrio and Streptococcus salivarius were depleted in CRC group compared to healthy group.At the same time,it was found that there were characteristic differences in the composition of gut microbiome among patients with CRC of different tumor location and pathological types.On average the WDIc scores of these 7 taxa was significantly higher in the CRC group than in the healthy group and positively correlated with tumor size and TNM staging.WDIc could differentiate CRC from the healthy samples with an AUC of 0.83 of the ROC curve,and with a sensitivity of 71.4%and a specificity of 83.1%.It had sensitivity of 71.0%and specificity of 87.0%when tested in the second cohort of patients and reasonable accuracy when further validated in the three public datasets.Conclusion:We developed an index to measure the gut microbiota dysbiosis in CRC patients,which had a good discriminatory power to distinguish CRC from the healthy samples,and may facilitate the potential clinical application of gut microbiota data in CRC detection.Part Ⅱ:The Correlation Between Changes of Gut Microbiome and Colorectal AdenomaObjective:Colorectal adenoma(CRA)is a classic precancerous disease of CRC and early diagnosis and treatment will greatly improve the prognosis of thus patients.Clinical screening indicators such as fecal occlusive blood test(FOBT)and serum carcinoma embryonic antigen(CEA)have low sensitivity to CRA,and there is still lack of non-invasive and efficient early detection methods.With the development of sequencing technology,changes in the abundance of gut microbiota in CRA patients have been found.This study aims to find a new marker to screen early colorectal lesions based on 16S rRNA sequencing.Methods:Stool samples were collected from 155 participants belonging to CRA group(96 cases,including advanced adenoma group 52 cases and nonadvanced adenoma group 44 cases)and healthy group(59 cases).FOBT and serum CEA were tested.Microbial compositions of the samples were assessed by 16S rRNA sequencing.The differences of intestinal microflora composition between control and CRA,advanced and non advanced adenoma patients were compared,and the changes of intestinal metagenome functions along the adenoma-carcinoma sequence were analyzed.Based on the relative abundance of CRA enriched or depleted taxa,a weighted dysbiosis index of adenoma(WDIa)model was built.The efficacy of WDIa,serum CEA and fecal OB were compared and analyzed.Results:Body mass index(BMI)of patients with adenoma was higher than healthy controls(P=0.014).And the size of intestinal polypus was much larger in patients with advanced adenoma than that of nonadvanced adenoma(P=9.15*10-18).Compared with control group,the abundance of Klebsiella Pneumoniae,Parabacteroides Merdae and Fusobacterium were significantly increased in the adenoma group.However,the abundance of Lachnospiraceae,Blautia,Actinobacteria,Bifidobacterium,Roseburia,Streptococcus,Pseudobutyrivibrio,Fusobacterium Ulcerans were significantly decreased.Prevotella and Bacteroides were significantly increased in nonadvanced adenoma group while Peptostreptococcus Anaerobius and Lachnoclostridium were significantly increased in advanced adenoma group.At the same time,30 Kyoto encyclopedia of genes and genomes(KEGG)pathways,such as amino acid metabolism,cardiovascular、nervous and endocrine system related diseases,peroxisome proliferators activated receptors(PPARs)and p53 pathways,were found to be differentially expressed in different stages of colorectal tumorigenesis.For FOBT,the sensitivity to CRA was 45.5%,the specificity was 91.5%,false positive rate was 8.5%and false negalive rate was 54.5%.For CEA,the sensitivity was 52.3%,the specificity was 70.6%,and the area under the ROC curve was 0.614.For WDIa,the sensitivity was 88.1%,the specificity was 54.2%,and the area under the ROC curve was 0.756.Conclusion:There were characteristic changes of intestinal microbial composition and metagenomic functions during the CRA stage of colorectal tumorigenesis.The sensitivity to CRA of WDIa in this study was significantly higher than that of FOBT and serum CEA.WDIa could identify early CRA more quickly and conveniently,and had the potential to be a noval marker of early colon lesions.Part Ⅲ:Alteration of the abundance of Parvimonas micra in the gut along the adenoma-carcinoma sequenceObjective:Parvimonas micra(P.micra)is reported to be associated with CRC.However,its association with CRA and its role in the initiation of colorectal tumors remain unknown,The present study aimed to clarify the relationship between P.micra and CRA and CRC by exploring the changes of P.micra abundance in an adenoma-carcinoma sequence in a SuZhou cohort and 4 public sequencing datasets.Methods:To investigate the alterations of P.micra abundance in the gut along the adenoma-carcinoma sequence,quantitative polymerase chain reaction(qPCR)was conducted to measure the relative abundance of P.micra in fecal samples from 277 subjects(128 patients with CRA,66 patients with CRC and 83 healthy individuals as controls)who underwent colonoscopy as outpatients.Then,the relative abundance of P.micra was analyzed in fecal samples from 596 subjects(185 healthy controls,158 CRC,253 CRA)in four public 16S rRNA sequencing datasets.Results:The qPCR results demonstrated that the CRA group had an abundance of P.micra(P=0.2)similar to that of the healthy group,while the CRC group had a significantly increased abundance(P=8.2×10-11).The level of P.micra effectively discriminated patients with CRC from healthy controls,while it poorly discriminated patients with CRA from healthy controls;with an area under the receiver operating characteristic curve of 0.867 for patients with CRC and 0.554 for patients with CRA.The same pattern of the alteration of P.micra abundance,which was low in healthy controls and patients with CRA but elevated in patients with CRC,was found in all four public sequencing datasets.Conclusion:These results suggested that P.micra was closely associated with CRC,and may serve as a diagnostic marker for CRC but not CRA.Moreover,it was indicated that P.micra,as an opportunistic pathogen of CRC,may promote CRC development but serve a limited role in CRA tumorigenesis.更多还原